title for essay about 1984

Home — Essay Samples — Literature — Books — 1984

Essays on 1984

Hook examples for "1984" essays, the dystopian warning hook.

Open your essay by discussing George Orwell's "1984" as a prophetic warning against totalitarianism and government surveillance. Explore how the novel's themes are eerily relevant in today's world.

The Orwellian Language Hook

Delve into the concept of Newspeak in "1984" and its parallels to modern language manipulation. Discuss how the novel's portrayal of controlled language reflects real-world instances of propaganda and censorship.

Big Brother is Watching Hook

Begin with a focus on surveillance and privacy concerns. Analyze the omnipresent surveillance in the novel and draw connections to contemporary debates over surveillance technologies, data privacy, and civil liberties.

The Power of Doublethink Hook

Explore the psychological manipulation in "1984" through the concept of doublethink. Discuss how individuals in the novel are coerced into accepting contradictory beliefs, and examine instances of cognitive dissonance in society today.

The Character of Winston Smith Hook

Introduce your readers to the protagonist, Winston Smith, and his journey of rebellion against the Party. Analyze his character development and the universal theme of resistance against oppressive regimes.

Technology and Control Hook

Discuss the role of technology in "1984" and its implications for control. Explore how advancements in surveillance technology, social media, and artificial intelligence resonate with the novel's themes of control and manipulation.

The Ministry of Truth Hook

Examine the Ministry of Truth in the novel, responsible for rewriting history. Compare this to the manipulation of information and historical revisionism in contemporary politics and media.

Media Manipulation and Fake News Hook

Draw parallels between the Party's manipulation of information in "1984" and the spread of misinformation and fake news in today's media landscape. Discuss the consequences of a distorted reality.

Relevance of Thoughtcrime Hook

Explore the concept of thoughtcrime and its impact on individual freedom in the novel. Discuss how society today grapples with issues related to freedom of thought, expression, and censorship.

The Importance of Fear in 1984

Life under the doublespeak office analysis, made-to-order essay as fast as you need it.

Each essay is customized to cater to your unique preferences

+ experts online

Big Brother in George Orwell’s "1984"

George orwell’s representation of authority as illustrated in his book, 1984, orwell's use of literary devices to portray the theme of totalitarianism in 1984, the culture of fear in 1984, a novel by george orwell, let us write you an essay from scratch.

• 450+ experts on 30 subjects ready to help
• Custom essay delivered in as few as 3 hours

1984 by George Orwell: Literary Devices to Portray Government Controlling Its Citizens

The use of language to control people in 1984, dictatorship of the people: orwell's 1984 as an allegory for the early soviet union, searching for truth in 1984, get a personalized essay in under 3 hours.

Expert-written essays crafted with your exact needs in mind

A World Without Love: The Ramifications of an Affectionless Society in 1984

On double-think and newspeak: orwell's language, the theme of survival and selfishness in the handmaid's tale in 1984, government surveillance in 1984 by george orwell: bogus security, george orwell's 1984 as a historical allegory, exploitation of language in george orwell's 1984, how orwell's 1984 is relevant to today's audience, the relation of orwel’s 1984 to the uighur conflict in china, symbolism in 1984: the soviet union as representation of the fears people, parallels to today in 1984 by george orwell, the relationship between power and emotions in 1984, proletariat vs protagonist: winston smith's class conflict in 1984, a review of george orwell’s book, 1984, o'brien as a dehumanizing villain in 1984, family in 1984 and persepolis, the philosophy of determinism in 1984, orwell's use of rhetorical strategies in 1984, control the citizens in the orwell's novel 1984, dangers of totalitarianism as depicted in 1984, dystopian life in '1984' was a real-life in china.

8 June 1949, George Orwell

Novel; Dystopia, Political Fiction, Social Science Fiction Novel

Winston Smith, Julia, O'Brien, Aaronson, Jones, and Rutherford, Ampleforth, Charrington, Tom Parsons, Syme, Mrs. Parsons, Katharine Smith

Since Orwell has been a democratic socialist, he has modelled his book and motives after the Stalinist Russia

Power, Repressive Behaviors, Totalitarianism, Mass Surveillance, Human Behaviors

The novel has brought up the "Orwellian" term, which stands for "Big Brother" "Thoughtcrime" and many other terms that we know well. It has been the reflection of totalitarianism

1984 represents a dystopian writing that has followed the life of Winston Smith who belongs to the "Party",which stands for the total control, which is also known as the Big Brother. It controls every aspect of people's lives. Is it ever possible to go against the system or will it take even more control. It constantly follows the fear and oppression with the surveillance being the main part of 1984. There is Party’s official O’Brien who is following the resistance movement, which represents an alternative, which is the symbol of hope.

Before George Orwell wrote his famous book, he worked for the BBC as the propagandist during World War II. The novel has been named 1980, then 1982 before finally settling on its name. Orwell fought tuberculosis while writing the novel. He died seven months after 1984 was published. Orwell almost died during the boating trip while he was writing the novel. Orwell himself has been under government surveillance. It was because of his socialist opinions. The slogan that the book uses "2 + 2 = 5" originally came from Communist Russia and stood for the five-year plan that had to be achieved during only four years. Orwell also used various Japanese propaganda when writing his novel, precisely his "Thought Police" idea.

“Who controls the past controls the future. Who controls the present controls the past.” “But if thought corrupts language, language can also corrupt thought.” “Being in a minority, even in a minority of one, did not make you mad. There was truth and there was untruth, and if you clung to the truth even against the whole world, you were not mad.” “Confession is not betrayal. What you say or do doesn't matter; only feelings matter. If they could make me stop loving you-that would be the real betrayal.” “Power is in tearing human minds to pieces and putting them together again in new shapes of your own choosing.” "But you could not have pure love or pure lust nowadays. No emotion was pure, because everything was mixed up with fear and hatred."

The most important aspect of 1984 is Thought Police, which controls every thought. It has been featured in numerous books, plays, music pieces, poetry, and anything that has been created when one had to deal with Social Science and Politics. Another factor that represents culmination is thinking about overthrowing the system or trying to organize a resistance movement. It has numerous reflections of the post WW2 world. Although the novella is graphic and quite intense, it portrays dictatorship and is driven by fear through the lens of its characters.

This essay topic is often used when writing about “The Big Brother” or totalitarian regimes, which makes 1984 a flexible topic that can be taken as the foundation. Even if you have to write about the use of fear by the political regimes, knowing the facts about this novel will help you to provide an example.

1. Enteen, G. M. (1984). George Orwell And the Theory of Totalitarianism: A 1984 Retrospective. The Journal of General Education, 36(3), 206-215. (https://www.jstor.org/stable/27797000) 2. Hughes, I. (2021). 1984. Literary Cultures, 4(2). (https://journals.ntu.ac.uk/index.php/litc/article/view/340) 3. Patai, D. (1982). Gamesmanship and Androcentrism in Orwell's 1984. PMLA, 97(5), 856-870. (https://www.cambridge.org/core/journals/pmla/article/abs/gamesmanship-and-androcentrism-in-orwells-1984/F1B026BE9D97EE0114E248AA733B189D) 4. Paden, R. (1984). Surveillance and Torture: Foucault and Orwell on the Methods of Discipline. Social Theory and Practice, 10(3), 261-271. (https://www.pdcnet.org/soctheorpract/content/soctheorpract_1984_0010_0003_0261_0272) 5. Tyner, J. A. (2004). Self and space, resistance and discipline: a Foucauldian reading of George Orwell's 1984. Social & Cultural Geography, 5(1), 129-149. (https://www.tandfonline.com/doi/abs/10.1080/1464936032000137966) 6. Kellner, D. (1990). From 1984 to one-dimensional man: Critical reflections on Orwell and Marcuse. Current Perspectives in Social Theory, 10, 223-52. (https://pages.gseis.ucla.edu/faculty/kellner/essays/from1984toonedimensional.pdf) 7. Samuelson, P. (1984). Good legal writing: of Orwell and window panes. U. Pitt. L. Rev., 46, 149. (https://heinonline.org/HOL/LandingPage?handle=hein.journals/upitt46&div=13&id=&page=) 8. Fadaee, E. (2011). Translation techniques of figures of speech: A case study of George Orwell's" 1984 and Animal Farm. Journal of English and Literature, 2(8), 174-181. (https://academicjournals.org/article/article1379427897_Fadaee.pdf) 9. Patai, D. (1984, January). Orwell's despair, Burdekin's hope: Gender and power in dystopia. In Women's Studies International Forum (Vol. 7, No. 2, pp. 85-95). Pergamon. (https://www.sciencedirect.com/science/article/abs/pii/0277539584900621) 10. Cole, M. B. (2022). The Desperate Radicalism of Orwell’s 1984: Power, Socialism, and Utopia in Dystopian Times. Political Research Quarterly, 10659129221083286. (https://journals.sagepub.com/doi/abs/10.1177/10659129221083286)

Relevant topics

• Between The World and Me
• Lord of The Flies
• A Rose For Emily
• The Tell Tale Heart
• The Alchemist
• Their Eyes Were Watching God
• Thank You Ma Am
• Frankenstein
• The Great Gatsby

By clicking “Check Writers’ Offers”, you agree to our terms of service and privacy policy . We’ll occasionally send you promo and account related email

No need to pay just yet!

Bibliography

We use cookies to personalyze your web-site experience. By continuing we’ll assume you board with our cookie policy .

• Instructions Followed To The Letter
• Deadlines Met At Every Stage
• Unique And Plagiarism Free

1984 Essay Topics & Examples

What can you say about the famous George Orwell’s book? With the 1984 essay topics and research titles gathered by our team , you’ll easily find the right words.

🏆 Best 1984 Essay Topics & Examples

📌 most interesting essay topics for 1984, 👍 good 1984 research paper topics, ❓ 1984 essay questions.

• George Orwell’s 1984: Winston and Julia’s Relationship Essay In the relationship, Julia teaches Winston the idea of love, and the love feeling is then manipulated and directed towards Big Brother.
• Historical Parallels Between George Orwell’s 1984 and Today Perhaps that is clearly illustrated by the quote that presupposes that whoever can control the past, has power to control the future; while whoever has the ability to control the present, wields the right to […]
• The Aspects of Human Nature That George Orwell Criticizes in His Work 1984 Compared to Today’s World The aspects of human nature that George Orwell criticizes in his work 1984 compared to today’s world Orwell in the novel 1984 represents the modern society be it capitalist or communist.
• Language in Orwell’s 1984 as a Means of Manipulation and Control One of the key themes in the novel is the control over language and rewriting history. Thus, it is apparent that control of language leads to the restriction of people’s feelings and thoughts.
• The Declaration of Independence and 1984 by George Orwell Another feature that relates the Declaration of Independence to 1984 is a demonstration of the tyranny of the ruler and the restriction of the citizen’s rights.
• George Orwell and Two of His Works “1984” and “Animal Farm” Orwell draws on his own personal experiences in the context of political terrorism to describe a life, lived in fear and guilt.
• Comparison of G. Orwell’s “1984”, R. Bradbury’s “Fahrenheit 451” and A. Huxley’s “Brave New World” The leadership is in charge of virtually each and every single activity that takes place in the lives of the inhabitants of the society.
• Dystopias “Brave New World” by Huxley and “1984” by Orwell The modern world is full of complications and the moments when it seems like a dystopia the darkest version of the future. In the novel, promiscuity is encouraged, and sex is a form of entertainment.
• The Dystopian Societies of “1984” and Brave New World The three features which are discussed in this respect are the division of the two societies into social strata, the use of state power and control over citizens, and the loss of people’s individualities.
• Two Opposite Worlds: “Utopia” and “1984” More criticizes the laws of the contemporary European society; he highlights that other countries, in the East for instance, have more fair laws; and after that he starts depicting Utopia, where all people live and […]
• Analysis of Enemy of the People and Nineteen Eighty Four Hovard evidences a good example of the barrier of doing the right things due to influences and the need to fulfill the desires of the people even if they are wrong.
• Literature Comparison: “One Flew Over the Cuckoo’s Nest” and “1984” It can be said that while both of these books address the issue of hidden methods of coercion, Nineteen-eighty Four provides a bleak vision of the future in which the whole of society is controlled […]
• Winston Smith, in the Novel “Nineteen Eighty-Four” Lastly, Winston Smith is not a hero, and individuals should not emulate and admire him as he is quick to surrender, indiscreet, and promotes the wealth of the ruling class.
• Unhappiness of Society in Orwell’s 1984 Dystopia His character is a strong individual who will not transgress the ideals of his party and is fully committed to him.
• Orwell’s 1984 Literary Analysis: Should the Majority Rule? The main character of the 1984 novel is Winston Smith, who is in his late 40s and who works in the Ministry of Truth or Minitruth, which is apparently the Ministry of Lies, since the […]
• Generation Z Through George Orwell’s “1984” Lens One of the things that the new generation lacks and that the old one had is respect for the opinion of an ideological opponent.
• “Novel 1984” by George Orwell The specific inspirations for the Oceania society from “1984” were The Soviet Union and Nazi Germany with their inherent propaganda, betrayal of the ideals of the revolution, concentration camps and misinformation.
• “Nineteen Eighty-Four” a Book by George Orwell The major purpose of the essay is to prove that, despite the wide-spread opinion of literary critics that the ideologies presented in the novel are all alike, it is still possible to indicate differences accounting […]
• Events in the 1984 by George Orwell This paper explores the similarities and dissimilarities between the book’s events and the occurrences of contemporary society in 2014. Orwell’s accounts in the book 1984 strike many similarities with the events happening in contemporary society.
• George Orwell’s Novel 1984 The world is involved in an endless war, and the political regime called Ingsoc and headed by a mystical Big Brother permanently looks for ways to control the citizens’ minds and private lives.
• Analysis of Books “Half the Sky How to Change the World”, “Gulliver’s Travel” and “1984” Comprehensively, the book Half the Sky How to Change the World exposes the rot that is human trafficking and tries to expose the severity of the trade and how it affects the world today.
• 1984 by George Orwell There are high hopes that the current settings of the twenty-first century and the predictable future of governance will be sustainable and responsible especially on issues of cultural identity and preservation.
• Understanding the Concept of Doublethink in the World of George Orwell’s “1984”
• The Weakness of Big Brother in “1984” by George Orwell
• The Theme of the Survival of a Hero in the Movie “Casablanca” and George Orwell’s “1984”
• The Truth About Communism and Totalitarism in George Orwell’s Novel “1984”
• The Similarities Between the Novels “Brave New World” by Aldous Huxley and “1984” by George Orwell
• Totalitarianism and Dystopia in George Orwell’s “1984”
• The Theme of History in “Brave New World” by Arthur Huxley and “1984” by George Orwell
• Theme Analysis in “Zeitoun” by Dave Eggers and “1984” by George Orwell
• The Philosophy of Determinism in “1984” by George Orwell
• The Power and Control of the Party in “1984” by George Orwell
• The Near Dystopian Future in a “Brave New World” by Aldous Huxley and “1984” by George Orwell
• The Suppression of Thoughts and the Elimination of Freedom in “1984” by George Orwell
• The Totalitarian Government of “1984” by George Orwell
• The Use of the Newspeak Language to Control and Manipulate in “1984” by George Orwell
• The Practice of Dehumanization by the Party in “1984” by George Orwell
• The Psychological Manipulation of Society in “1984” by George Orwell
• Theme of Betrayal in the Novel “1984” by George Orwell
• The Roles of Love, Government, Freedom, Education, and Pleasure in George Orwell’s “1984”
• The Idea of Humans Being Naturally Rebellious in “1984” by George Orwell
• The World of Deceit and Propaganda in George Orwell’s “1984”
• The Importance of Winston and Julie’s Romantic Relationship in George Orwell’s “1984”
• The Inferiority of Women in “Brave New World” by Aldous Huxley and “1984” by George Orwell
• The Utopian Society in “1984” by George Orwell
• The Significance of the Elements of Political Protest in “1984” by George Orwell
• The Necessities for a Dystopian Society in George Orwell’s “1984” and Its Possibility in the Modern Era
• The Role of Newspeak in the Inner Party’s Philosophy and Propaganda in “1984” by George Orwell
• Totalitarian Society in George Orwell’s “1984”
• The Mirrored Worlds in Novels “1984” by George Orwell and “The Handmaid’s Tale” by Margaret Atwood
• Totalitarian Goverments in George Orwell’s “1984”
• The Pleasure Principle in “Brave New World” by Aldous Huxley and “1984” by George Orwell
• The Parallelism of Today’s Society to the Social Conditions Found in George Orwell’s “1984”
• Winston Smith in the Novel “1984” by George Orwell
• The Three Important Aspects of the Fictional World in “1984” by George Orwell
• The Verbal and Situation Irony in George Orwell’s “1984”
• Understanding Dystopia in “1984” by George Orwell and “The Handmaid’s Tale” by Margaret Atwood
• The Government’s Suppression of Freedom in “1984” by George Orwell
• The Influence of Stalinist Russia’s Total Control, Censorship, and Terror on George Orwell’s “1984”
• The Opening of Public Opinions to Future World in George Orwell’s “1984”
• The Political Satire of the Novel “1984” by George Orwell
• Triumph and Futility in “The Fountainhead” by Ayn Rand and “1984” by George Orwell
• The Exploration of Truth and Reality in “1984” by George Orwell
• The Societal Impact of Surveillance and the “Big Brother” Concept in “1984” by George Orwell
• The Traits of Society in George Orwell’s “1984”
• The Use and Abuse of Power in “1984” by George Orwell
• The Themes of the Dangers of Psychological Manipulation and Physical Control in “1984” by George Orwell
• The Impact of the Advances in Technology in “1984” by George Orwell
• The Understanding and Manipulation of Emotion as a Tool for Building Power in “1984” by George Orwell
• The Use of Foreshadowing in George Orwell’s “1984”
• The Government’s Attempt to Control Citizen’s Minds and Bodies in George Orwell’s “1984”
• The Four Essential Freedoms and the Freedom of Fear in “1984” by George Orwell
• How Does the George Orwell Use Language to Create a Sense of Place in “1984”?
• What Is the Significance of Coffee in “1984”?
• Why Did Winston Betray Julia in “1984”?
• What Role Does Contradiction Serve Within the Framework of Doublethink in “1984”?
• How Does “1984” Relate to Dystopian Literature?
• Is There Evidence in “1984” That Supports the Poster That Says “Big Brother Is Watching You”?
• What Was the Two Minutes Hate in “1984”?
• How Does Winston View His Job at the Ministry of Truth in “1984”?
• Why Is Winston So Afraid of Rats in “1984”?
• How Does “1984” Relate to Contemporary Politics and Society?
• How Is Free Will Seen in George Orwell’s ‘’1984’’?
• How Does the Interaction of Text and Reader Create Meaning in the Novel “1984” by George Orwell?
• What Is the Role of Women in “1984”?
• How Do Winston and Julia Differ in Their Views of the Past in “1984”?
• How Is Technology Used to Control the Citizens in “1984”?
• How Does the Party Use Propaganda in “1984”?
• What Are the Morals and Ethical Views of Winston and Julia in the Novel “1984”?
• What Does the Rat Symbolize in “1984”?
• How Are “1984” and “Harrison Bergeron” Alike and Different?
• What Does Memory Hole Mean in “1984”?
• What Is the Purpose of the Record’s Department in “1984”?
• Why Does the Party Discourage Romantic Relationships Between Party Members in “1984”?
• What Was Julia’s Room 101 in “1984”?
• How Does George Orwell Reveal Character in “1984”?
• What Warnings Can We Take From Orwell’s “1984”?
• How Are Characters Brainwashed in “1984”?
• How Effectively Does Orwell Introduce the Reader to the New Society of “1984” in Chapter One of the Novel?
• What Is the Significance of the Name Ministry of Love in “1984”?
• What Is the Main Problem in “1984”?
• What Is O’Brien’s Vision for the Future of Oceania in “1984”?
• Call of the Wild Questions
• The House on Mango Street Research Topics
• The Alchemist Questions
• The Awakening Questions
• The Great Gatsby Ideas
• A Midsummer Night’s Dream Titles
• The Glass Menagerie Paper Topics
• The Handmaid’s Tale Research Ideas
• Chicago (A-D)
• Chicago (N-B)

IvyPanda. (2024, February 20). 1984 Essay Topics & Examples. https://ivypanda.com/essays/topic/1984-essay-examples/

"1984 Essay Topics & Examples." IvyPanda , 20 Feb. 2024, ivypanda.com/essays/topic/1984-essay-examples/.

IvyPanda . (2024) '1984 Essay Topics & Examples'. 20 February.

IvyPanda . 2024. "1984 Essay Topics & Examples." February 20, 2024. https://ivypanda.com/essays/topic/1984-essay-examples/.

1. IvyPanda . "1984 Essay Topics & Examples." February 20, 2024. https://ivypanda.com/essays/topic/1984-essay-examples/.

Bibliography

IvyPanda . "1984 Essay Topics & Examples." February 20, 2024. https://ivypanda.com/essays/topic/1984-essay-examples/.

George Orwell

Ask litcharts ai: the answer to your questions.

Welcome to the LitCharts study guide on George Orwell's 1984 . Created by the original team behind SparkNotes, LitCharts are the world's best literature guides.

1984: Introduction

1984: plot summary, 1984: detailed summary & analysis, 1984: themes, 1984: quotes, 1984: characters, 1984: symbols, 1984: theme wheel, brief biography of george orwell.

Historical Context of 1984

Other books related to 1984.

• Full Title: Nineteen Eighty-Four: A Novel
• When Written: 1945-49; outline written 1943
• Where Written: Jura, Scotland
• When Published: June 1949
• Literary Period: Late Modernism
• Genre: Novel / Satire / Parable
• Setting: London in the year 1984
• Climax: Winston is tortured in Room 101
• Antagonist: O'Brien
• Point of View: Third-Person Limited

Extra Credit for 1984

Outspoken Anti-Communist. Orwell didn't just write literature that condemned the Communist state of the USSR. He did everything he could, from writing editorials to compiling lists of men he knew were Soviet spies, to combat the willful blindness of many intellectuals in the West to USSR atrocities.

Working Title. Orwell's working title for the novel was The Last Man in Europe .

Doublethink Is Stronger Than Orwell Imagined

What 1984 means today

No novel of the past century has had more influence than George Orwell’s 1984 . The title, the adjectival form of the author’s last name, the vocabulary of the all-powerful Party that rules the superstate Oceania with the ideology of Ingsoc— doublethink , memory hole , unperson , thoughtcrime , Newspeak , Thought Police , Room 101 , Big Brother —they’ve all entered the English language as instantly recognizable signs of a nightmare future. It’s almost impossible to talk about propaganda, surveillance, authoritarian politics, or perversions of truth without dropping a reference to 1984. Throughout the Cold War, the novel found avid underground readers behind the Iron Curtain who wondered, How did he know?

It was also assigned reading for several generations of American high-school students. I first encountered 1984 in 10th-grade English class. Orwell’s novel was paired with Aldous Huxley’s Brave New World , whose hedonistic and pharmaceutical dystopia seemed more relevant to a California teenager in the 1970s than did the bleak sadism of Oceania. I was too young and historically ignorant to understand where 1984 came from and exactly what it was warning against. Neither the book nor its author stuck with me. In my 20s, I discovered Orwell’s essays and nonfiction books and reread them so many times that my copies started to disintegrate, but I didn’t go back to 1984 . Since high school, I’d lived through another decade of the 20th century, including the calendar year of the title, and I assumed I already “knew” the book. It was too familiar to revisit.

Read: Teaching ‘1984’ in 2016

So when I recently read the novel again, I wasn’t prepared for its power. You have to clear away what you think you know, all the terminology and iconography and cultural spin-offs, to grasp the original genius and lasting greatness of 1984 . It is both a profound political essay and a shocking, heartbreaking work of art. And in the Trump era , it’s a best seller .

The Ministry of Truth: The Biography of George Orwell’s 1984 , by the British music critic Dorian Lynskey, makes a rich and compelling case for the novel as the summation of Orwell’s entire body of work and a master key to understanding the modern world. The book was published in 1949, when Orwell was dying of tuberculosis , but Lynskey dates its biographical sources back more than a decade to Orwell’s months in Spain as a volunteer on the republican side of the country’s civil war. His introduction to totalitarianism came in Barcelona, when agents of the Soviet Union created an elaborate lie to discredit Trotskyists in the Spanish government as fascist spies.

Left-wing journalists readily accepted the fabrication, useful as it was to the cause of communism. Orwell didn’t, exposing the lie with eyewitness testimony in journalism that preceded his classic book Homage to Catalonia —and that made him a heretic on the left. He was stoical about the boredom and discomforts of trench warfare—he was shot in the neck and barely escaped Spain with his life—but he took the erasure of truth hard. It threatened his sense of what makes us sane, and life worth living. “History stopped in 1936,” he later told his friend Arthur Koestler, who knew exactly what Orwell meant. After Spain, just about everything he wrote and read led to the creation of his final masterpiece. “History stopped,” Lynskey writes, “and Nineteen Eighty-Four began.”

The biographical story of 1984 —the dying man’s race against time to finish his novel in a remote cottage on the Isle of Jura , off Scotland—will be familiar to many Orwell readers. One of Lynskey’s contributions is to destroy the notion that its terrifying vision can be attributed to, and in some way disregarded as, the death wish of a tuberculosis patient. In fact, terminal illness roused in Orwell a rage to live—he got remarried on his deathbed—just as the novel’s pessimism is relieved, until its last pages, by Winston Smith’s attachment to nature, antique objects, the smell of coffee, the sound of a proletarian woman singing, and above all his lover, Julia. 1984 is crushingly grim, but its clarity and rigor are stimulants to consciousness and resistance. According to Lynskey, “Nothing in Orwell’s life and work supports a diagnosis of despair.”

Lynskey traces the literary genesis of 1984 to the utopian fictions of the optimistic 19th century—Edward Bellamy’s Looking Backward (1888); the sci-fi novels of H. G. Wells, which Orwell read as a boy—and their dystopian successors in the 20th, including the Russian Yevgeny Zamyatin’s We (1924) and Huxley’s Brave New World (1932). The most interesting pages in The Ministry of Truth are Lynskey’s account of the novel’s afterlife. The struggle to claim 1984 began immediately upon publication, with a battle over its political meaning. Conservative American reviewers concluded that Orwell’s main target wasn’t just the Soviet Union but the left generally. Orwell, fading fast, waded in with a statement explaining that the novel was not an attack on any particular government but a satire of the totalitarian tendencies in Western society and intellectuals: “The moral to be drawn from this dangerous nightmare situation is a simple one: Don’t let it happen. It depends on you .” But every work of art escapes the artist’s control—the more popular and complex, the greater the misunderstandings.

Lynskey’s account of the reach of 1984 is revelatory. The novel has inspired movies, television shows, plays, a ballet, an opera, a David Bowie album , imitations, parodies, sequels, rebuttals, Lee Harvey Oswald, the Black Panther Party, and the John Birch Society. It has acquired something of the smothering ubiquity of Big Brother himself: 1984 is watching you. With the arrival of the year 1984, the cultural appropriations rose to a deafening level. That January an ad for the Apple Macintosh was watched by 96 million people during the Super Bowl and became a marketing legend. The Mac, represented by a female athlete, hurls a sledgehammer at a giant telescreen and explodes the shouting face of a man—oppressive technology—to the astonishment of a crowd of gray zombies. The message: “You’ll see why 1984 won’t be like ‘1984.’ ”

The argument recurs every decade or so: Orwell got it wrong. Things haven’t turned out that bad. The Soviet Union is history. Technology is liberating. But Orwell never intended his novel to be a prediction, only a warning. And it’s as a warning that 1984 keeps finding new relevance. The week of Donald Trump’s inauguration, when the president’s adviser Kellyanne Conway justified his false crowd estimate by using the phrase alternative facts , the novel returned to the best-seller lists. A theatrical adaptation was rushed to Broadway. The vocabulary of Newspeak went viral. An authoritarian president who stood the term fake news on its head, who once said, “What you’re seeing and what you’re reading is not what’s happening,” has given 1984 a whole new life.

What does the novel mean for us? Not Room 101 in the Ministry of Love, where Winston is interrogated and tortured until he loses everything he holds dear. We don’t live under anything like a totalitarian system. “By definition, a country in which you are free to read Nineteen Eighty-Four is not the country described in Nineteen Eighty-Four ,” Lynskey acknowledges. Instead, we pass our days under the nonstop surveillance of a telescreen that we bought at the Apple Store, carry with us everywhere, and tell everything to, without any coercion by the state. The Ministry of Truth is Facebook, Google, and cable news. We have met Big Brother and he is us.

Trump’s election brought a rush of cautionary books with titles like On Tyranny , Fascism: A Warning , and How Fascism Works . My local bookstore set up a totalitarian-themed table and placed the new books alongside 1984 . They pointed back to the 20th century—if it happened in Germany, it could happen here—and warned readers how easily democracies collapse. They were alarm bells against complacency and fatalism—“ the politics of inevitability ,” in the words of the historian Timothy Snyder, “a sense that the future is just more of the present, that the laws of progress are known, that there are no alternatives, and therefore nothing really to be done.” The warnings were justified, but their emphasis on the mechanisms of earlier dictatorships drew attention away from the heart of the malignancy—not the state, but the individual. The crucial issue was not that Trump might abolish democracy but that Americans had put him in a position to try. Unfreedom today is voluntary. It comes from the bottom up.

We are living with a new kind of regime that didn’t exist in Orwell’s time. It combines hard nationalism—the diversion of frustration and cynicism into xenophobia and hatred—with soft distraction and confusion: a blend of Orwell and Huxley, cruelty and entertainment. The state of mind that the Party enforces through terror in 1984 , where truth becomes so unstable that it ceases to exist, we now induce in ourselves. Totalitarian propaganda unifies control over all information, until reality is what the Party says it is—the goal of Newspeak is to impoverish language so that politically incorrect thoughts are no longer possible. Today the problem is too much information from too many sources, with a resulting plague of fragmentation and division—not excessive authority but its disappearance, which leaves ordinary people to work out the facts for themselves, at the mercy of their own prejudices and delusions.

During the 2016 U.S. presidential campaign, propagandists at a Russian troll farm used social media to disseminate a meme: “ ‘The People Will Believe What the Media Tells Them They Believe.’  — George Orwell.” But Orwell never said this. The moral authority of his name was stolen and turned into a lie toward that most Orwellian end: the destruction of belief in truth. The Russians needed partners in this effort and found them by the millions, especially among America’s non-elites. In 1984 , working-class people are called “proles,” and Winston believes they’re the only hope for the future. As Lynskey points out, Orwell didn’t foresee “that the common man and woman would embrace doublethink as enthusiastically as the intellectuals and, without the need for terror or torture, would choose to believe that two plus two was whatever they wanted it to be.”

We stagger under the daily load of doublethink pouring from Trump, his enablers in the Inner Party, his mouthpieces in the Ministry of Truth, and his fanatical supporters among the proles. Spotting doublethink in ourselves is much harder. “To see what is in front of one’s nose needs a constant struggle,” Orwell wrote . In front of my nose, in the world of enlightened and progressive people where I live and work, a different sort of doublethink has become pervasive. It’s not the claim that true is fake or that two plus two makes five. Progressive doublethink—which has grown worse in reaction to the right-wing kind—creates a more insidious unreality because it operates in the name of all that is good. Its key word is justice —a word no one should want to live without. But today the demand for justice forces you to accept contradictions that are the essence of doublethink.

For example, many on the left now share an unacknowledged but common assumption that a good work of art is made of good politics and that good politics is a matter of identity. The progressive view of a book or play depends on its political stance, and its stance—even its subject matter—is scrutinized in light of the group affiliation of the artist: Personal identity plus political position equals aesthetic value. This confusion of categories guides judgments all across the worlds of media, the arts, and education, from movie reviews to grant committees. Some people who register the assumption as doublethink might be privately troubled, but they don’t say so publicly. Then self-censorship turns into self-deception, until the recognition itself disappears—a lie you accept becomes a lie you forget. In this way, intelligent people do the work of eliminating their own unorthodoxy without the Thought Police.

Recommended Reading

A lost scottish island, george orwell, and the future of maps.

David Simon and E.L. Doctorow on 'the Potential for the Orwellian Nightmare'

Surviving the Worst Skydiving Accident in History

Orthodoxy is also enforced by social pressure, nowhere more intensely than on Twitter, where the specter of being shamed or “canceled” produces conformity as much as the prospect of adding to your tribe of followers does. This pressure can be more powerful than a party or state, because it speaks in the name of the people and in the language of moral outrage, against which there is, in a way, no defense. Certain commissars with large followings patrol the precincts of social media and punish thought criminals, but most progressives assent without difficulty to the stifling consensus of the moment and the intolerance it breeds—not out of fear, but because they want to be counted on the side of justice.

This willing constriction of intellectual freedom will do lasting damage. It corrupts the ability to think clearly, and it undermines both culture and progress. Good art doesn’t come from wokeness, and social problems starved of debate can’t find real solutions. “Nothing is gained by teaching a parrot a new word,” Orwell wrote in 1946. “What is needed is the right to print what one believes to be true, without having to fear bullying or blackmail from any side.” Not much has changed since the 1940s. The will to power still passes through hatred on the right and virtue on the left.

1984 will always be an essential book, regardless of changes in ideologies, for its portrayal of one person struggling to hold on to what is real and valuable. “Sanity is not statistical,” Winston thinks one night as he slips off to sleep. Truth, it turns out, is the most fragile thing in the world. The central drama of politics is the one inside your skull.

This article appears in the July 2019 print edition with the headline “George Orwell’s Unheeded Warning.”

​When you buy a book using a link on this page, we receive a commission. Thank you for supporting The Atlantic.

1984 George Orwell

1984 essays are academic essays for citation. These papers were written primarily by students and provide critical analysis of 1984 by George Orwell.

1984 Material

• Study Guide
• Lesson Plan

Join Now to View Premium Content

GradeSaver provides access to 2360 study guide PDFs and quizzes, 11007 literature essays, 2767 sample college application essays, 926 lesson plans, and ad-free surfing in this premium content, “Members Only” section of the site! Membership includes a 10% discount on all editing orders.

1984 Essays

The power of thought as tangible resistance anonymous 12th grade.

In George Orwell’s renowned novel, 1984, the protagonist, Winston Smith continues to preserve his normal, day to day tendencies while secretly questioning the rigid policies of Oceania’s ominously dark society privately within his mind. Although...

The Reflection of George Orwell Crystal Epps

"On each landing, opposite the lift shaft, the poster with the enormous face gazed from the wall. It was one of those pictures which are so contrived that the eyes follow you about when you move. BIG BROTHER IS WATCHING YOU, the caption beneath it...

Totalitarian Collectivism in 1984, or, Big Brother Loves You Tiffany Shropshire

Following the political upheaval and struggle for power after the second world war, George Orwell's novel 1984 cautions against the dangers of oppression and exemplifies the consequential nightmarish world of the near future. The plot traces the...

Sex as Rebellion Joe Ward

The opening of Book Two of 1984, in which Winston meets Julia and begins the erotic affair he has so deeply desired, commences the main section of the novel and strikes an immediate contrast between the two lovers. Unlike Winston, Julia is neither...

Class Ties: The Dealings of Human Nature Depicted through Social Classes in 1984 Zachary Zill

In George Orwell's 1984, the differences and relationships between the proles, the Outer Party, and the Inner Party reflect different aspects of human nature and the various levels of the human psyche. The most base, savage level of humanity is...

1984: The Ultimate Parody of the Utopian World Anonymous

"When Thomas More wrote Utopia in 1515, he started a literary genre with lasting appeal for writers who wanted not only to satirize existing evils but to postulate the state, a kind of Golden Age in the face of reality" (Hewitt 127). Unlike a...

Class Conflict: Winston Smith in George Orwell's 1984 Sarah Standish

The title year of George Orwell's most famous novel is nineteen years past, but the dystopian vision it draws has retained its ability to grip readers with a haunting sense of foreboding about the future. At the heart of many of the issues touched...

Methods of Control in 1984 and Brave New World Anonymous

The difference between the methods of control in 1984 and BRAVE NEW WORLD is the difference between external control by force and internal control, enforced only by the citizen's own mind. While 1984's method has real-world precedent and seems...

Time in Modernist Literature Nathan Ragolia

Perception of time represents a major motif in modernist literature. Many works address the subjectivity of our experiences, including how we process and consider the passage of time. Due to the modernist and post-modernist emphasis on style and...

The Impossibility of Redemption for Winston Smith in 1984 Timothy Sexton

In George Orwell's 1984, Winston Smith cannot escape the state's domination. Yet his inability is not only because of government power. Rather, even if he did have an opportunity to leave Oceania, his actions indicate that he would not have the...

Selfishness and Survival in The Handmaid’s Tale and 1984 Soh Li Yin

Are Winston, Julia and Offred eventually made into ‘reluctantly-selfish’ victims of totalitarian regimes or are they innately ‘pragmatically-selfish’ beings? Discuss in relation to The Handmaid’s Tale and 1984.

Offred and Winston, the main...

Power and Emotion in Orwell’s 1984 Anonymous

“How does one man assert his power over another, Winston?” O’Brien asks. Winston’s answer: “By making him suffer” (214). These two characters inhabit George Orwell’s vision of a future totalitarian government that has evolved to its most...

Imagery of Totalitarianism in Nineteen Eighty-Four Elizabeth Marcil 11th Grade

In the novel Nineteen Eighty-Four, Orwell uses several literary techniques to develop the theme that totalitarianism is destructive. He does so by using extensive imagery, focusing on the deterioration of the Victory Mansions, the canteen where...

Pursuit of Truth in 1984 Anonymous College

Contemporary political discourse often references George Orwell’s 1984 as an example of how government interference infringes on our rights as individuals while we remain complacent in the face of these violations. For example, the falsification...

Victorian, Romantic and Modernist Literature: Style as Cultural Commentary Anonymous College

Tony Harrison’s “A Cold Coming,” William Wordsworth’s “Lines Composed a Few Miles Above Tintern Abbey,” Emily Bronte’s Wuthering Heights and George Orwell’s 1984 each display distinct sensibilities that reflect the time from which they emerged....

The Currency of Power in 1984 Katherine Knapp College

The power of words is enough to control an entire nation. Although many would consider physical power and brute force to be absolute power, George Orwell’s 1984 demonstrates a dystopian society where language is the ultimate form of power. The...

Orwell's Language: Thought Control Tom Armstrong College

George Orwell’s 1984 portrays a dystopian society whose values and freedoms have been marred through the manipulation of language and thus thought processes. Language has become a tool of mind control for the oppressive government and...

The Freedom to Be Dominated: A Historical Comparison of 1984 to Communist Russia Anonymous 11th Grade

A government of an ideal society is meant to represent the people. It is the people’s choice to support, to select, and to seize government. The idea of open communication is employed as a way for people to choose the best representative. With the...

Models of Rebellion in 1984 and V for Vendetta Joseph Latorcai 12th Grade

Problems faced by characters in literature often repeat themselves, and when these characters decide to solve these standard problems, their actions are often more similar than they first appear. This idea is evident when comparing the actions...

Freud's Impact on 1984 Anonymous College

In his treatise Civilization and Its Discontents , Freud makes an interesting statement about advanced society. He argues that “the price of progress in civilization is paid in forfeiting happiness through the heightening of the sense of guilt,” to...

O’Brien’s Moral Dehumanization: Villainy in "1984" Dylan Kostadinov 10th Grade

“Nobody is a villain in their own story. We're all the heroes of our own stories.” According to George R.R. Martin, an estimable American novelist, an individual's perspective ultimately decides whether he views himself as a protagonist and deems...

Rebellion Across Media: Analyzing "1984" and "Metropolis" Joonhwy Kwon 12th Grade

George Orwell’s 1984 (1949) is a cautionary novel which explores a dystopian society mired in propaganda and totalitarianism. Similarly, director Fritz Lang’s Metropolis (1927) is a critique of a futuristic world where growth and industralisation...

Totalitarian Techniques in 1984 and Red Azalea Anonymous 10th Grade

In order for one to exist in a totalitarian society whose government is successful in its control, one must deal on a day-to-day basis with strong persuasion and propaganda. These totalitarian societies have an iron grip on their people, leaving...

Humanity's Fear: A Comparison of 1984 and Metropolis Anonymous 12th Grade

The fear of a dystopian future that is explored in both Fritz Lang’s film Metropolis and George Orwell’s novel Nineteen Eighty Four is reflective of the values of the societies at the time and the context of the authors. As authors are considered...

A Summary and Analysis of George Orwell’s Nineteen Eighty-Four

By Dr Oliver Tearle (Loughborough University)

George Orwell’s Nineteen Eighty-Four , completed in 1948 and published a year later, is a classic example of dystopian fiction. Indeed, it’s surely the most famous dystopian novel in the world, even if its ideas are known by far more people than have actually read it. (According to at least one survey , Nineteen Eighty-Four is the book people most often claim to have read when they haven’t.)

Like many novels that are more known about than are carefully read and analysed, Nineteen Eighty-Four is actually a more complex work than the label ‘nightmare dystopian vision’ can convey. Before we offer an analysis of the novel’s themes and origins, let’s briefly recap the plot.

Nineteen Eighty-Four : plot summary

In the year 1984, Britain has been renamed Airstrip One and is a province of Oceania, a vast totalitarian superstate ruled by ‘the Party’, whose politics are described as Ingsoc (‘English Socialism’). Big Brother is the leader of the Party, which keeps its citizens in a perpetual state of fear and submission through a variety of means.

Surveillance is a key part of the novel’s world, with hidden microphones (which are found in the countryside as well as urban areas, and can identify not only what is said but also who says it) and two-way telescreen monitors being used to root out any dissidents, who disappear from society with all trace of their existence wiped out.

They become, in the language of Newspeak (the language used by people in the novel), ‘unpersons’. People are short of food, perpetually on the brink of starvation, and going about in fear for their lives.

The novel’s setting is London, where Trafalgar Square has been renamed Victory Square and the statue of Horatio Nelson atop Nelson’s Column has been replaced by one of Big Brother. Through such touches, Orwell defamiliarises the London of the 1940s which the original readers would have recognised, showing how the London they know might be transformed under a totalitarian regime.

The novel’s protagonist is Winston Smith, who works at the Ministry of Truth, rewriting historical records so they are consistent with the state’s latest version of history. However, even though his day job involves doing the work of the Party, Winston longs to escape the oppressive control of the Party, hoping for a rebellion.

Winston meets the owner of an antique shop named Mr Charrington, from whom he buys a diary in which he can record his true feelings towards the Party. Believing the working-class ‘proles’ are the key to a revolution, Winston visits them, but is disappointed to find them wholly lacking in any political understanding.

Meanwhile, hearing of the existence of an underground resistance movement known as the Brotherhood – which has been formed by the rival of Big Brother, a man named Emmanuel Goldstein – Winston suspects that O’Brien, who also works with him, is involved with this resistance.

At lunch with another colleague, named Syme, Winston learns that the English language is being rewritten as Newspeak so as to control and influence people’s thought, the idea being that if the word for an idea doesn’t exist in the language, people will be unable to think about it.

Winston meets a woman named Julia who works for the Ministry of Truth, maintaining novel-writing machines, but believes she is a Party spy sent to watch him. But then Julia passes a clandestine love message to him and the two begin an affair – which is itself illicit since the Party decrees that sex is for reproduction alone, rather than pleasure.

We gradually learn more about Winston’s past, including his marriage to Katherine, from whom he is now separated. Syme, who had been working on Newspeak, disappears in mysterious circumstances: something Winston had predicted.

O’Brien invites Winston to his flat, declaring himself – as Winston had also predicted – a member of the Brotherhood, the resistance against the Party. He gives Winston a copy of the book written by Goldstein, the leader of the Brotherhood.

When Oceania’s enemy changes during the ritual Hate Week, Winston is tasked with making further historical revisions to old newspapers and documents to reflect this change.

Meanwhile, Winston and Julia secretly read Goldstein’s book, which explains how the Party maintains its totalitarian power. As Winston had suspected, the secret to overthrowing the Party lies in the vast mass of the population known as the ‘proles’ (derived from ‘proletarian’, Marx’s term for the working classes). It argues that the Party can be overthrown if proles rise up against it.

But shortly after this, Winston and Julia are arrested, having been shopped to the authorities by Mr Charrington (whose flat above his shop they had been using for their illicit meetings). It turns out that both he and O’Brien work for the Thought Police, on behalf of the Party.

At the Ministry of Love, O’Brien tells Winston that Goldstein’s book was actually written by him and other Party members, and that the Brotherhood may not even exist. Winston endures torture and starvation in an attempt to grind him down so he will accept Big Brother.

In Room 101, a room in which a prisoner is exposed to their greatest fear, Winston is placed in front of a wire cage containing rats, which he fears above all else. Winston betrays Julia, wishing she could take his place and endure this suffering instead.

His reprogramming complete, Winston is allowed to go free, but he is essentially living under a death sentence: he knows that one day he will be summoned by the authorities and shot for his former treachery.

He meets Julia one day, and learns that she was subjected to torture at the Ministry of Love as well. They have both betrayed each other, and part ways. The novel ends with Winston accepting, after all, that the Party has won and that ‘he loved Big Brother.’

Nineteen Eighty-Four : analysis

Nineteen Eighty-Four is probably the most famous novel about totalitarianism, and about the dangers of allowing a one-party state where democracy, freedom of movement, freedom of speech, and even freedom of thought are all outlawed. The novel is often analysed as a warning about the dangers of allowing a creeping totalitarianism into Britain, after the horrors of such regimes in the Soviet Union, Nazi Germany, and elsewhere had been witnessed.

Because of this quality of the book, it is often called ‘prophetic’ and a ‘nightmare vision of the future’, among other things.

However, books set in the future are rarely simply about the future. They are not mere speculation, but are grounded in the circumstances in which they were written.

Indeed, we might go so far as to say that most dystopian novels, whilst nominally set in an imagined future, are really using their future setting to reflect on what are already firmly established social or political ideas. In the case of Orwell and Nineteen Eighty-Four , this means the novel reflects the London of the 1940s.

By the time he came to write the novel, Orwell already had a long-standing interest in using his writing to highlight the horrors of totalitarianism around the world, especially following his experience fighting in the Spanish Civil War in the 1930s. As Orwell put it in his essay ‘ Why I Write ’, all of his serious work written since 1936 was written ‘ against totalitarianism and for democratic socialism’.

In his analysis of Nineteen Eighty-Four in his study of Orwell, George Orwell (Reader’s Guides) , Jeffrey Meyers argues convincingly that, rather than being a nightmare vision of the future, a prophetic or speculative work, Orwell’s novel is actually a ‘realistic synthesis and rearrangement of familiar materials’ – indeed, as much of Orwell’s best work is.

His talent lay not in original imaginative thinking but in clear-headed critical analysis of things as they are: his essays are a prime example of this. Nineteen Eighty-Four is, in Meyer’s words, ‘realistic rather than fantastic’.

Indeed, Orwell himself stated that although the novel was ‘in a sense a fantasy’, it is written in the form of the naturalistic novel, with its themes and ideas having been already ‘partly realised in Communism and fascism’. Orwell’s intention, as stated by Orwell himself, was to take the totalitarian ideas that had ‘taken root’ in the minds of intellectuals all over Europe, and draw them out ‘to their logical consequences’.

Like much classic speculative fiction – the novels and stories of J. G. Ballard offer another example – the futuristic vision of the author is more a reflection of contemporary anxieties and concerns. Meyers goes so far as to argue that Nineteen Eighty-Four is actually the political regimes of Nazi Germany and Stalinist Russia ‘transposed’ into London of the early 1940s, during the Second World War.

Certainly, many of the most famous features of Nineteen Eighty-Four were suggested to Orwell by his time working at the BBC in London in the first half of the 1940s: it is well-known that the Ministry of Truth was based on the bureaucratic BBC with its propaganda department, while the infamous Room 101 was supposedly named after a room of that number in the BBC building, in which Orwell had to endure tedious meetings.

The technology of the novel, too, was familiar by the 1940s, involving little innovation or leaps of imagination from Orwell (‘telescreens’ being a natural extension of the television set: BBC TV had been established in 1936, although the Second World War pushed back its development somewhat).

Orwell learned much about the workings of Stalinism from reading Trotsky’s The Revolution Betrayed (1937), written by one of the leading figures in the Russian Revolution of 1917 who saw Stalinist Russia as the antithesis of what Trotsky, Lenin, and those early revolutionaries had been striving to achieve. (This would also be important for Orwell’s Animal Farm , of course.)

And indeed, many of the details surrounding censorship – the rewriting of history, the suppression of dissident literature, the control of the language people use to express themselves and even to think in – were also derived from Orwell’s reading of life in Soviet Russia. Surveillance was also a key element of the Stalinist regime, as in other Communist countries in Europe.

The moustachioed figure of Big Brother in Nineteen Eighty-Four recalls nobody so much as Josef Stalin himself. Not only the ideas of ‘thought crime’ and ‘thought police’, but even the terms themselves, predate Orwell’s use of them: they were first recorded in a 1934 book about Japan.

One of the key questions Winston asks himself in Nineteen Eighty-Four is what the Party is trying to achieve. O’Brien’s answer is simple: the maintaining of power for its own sake. Many human beings want to control other human beings, and they can persuade a worrying number of people to go along with their plans and even actively support them.

Despite the fact that they are starving and living a miserable life, many of the people in Airstrip One love Big Brother, viewing him not as a tyrannical dictator but as their ‘Saviour’ (as one woman calls him). Again, this detail was taken from accounts of Stalin, who was revered by many Russians even though they were often living a wretched life under his rule.

Another key theme of Orwell’s novel is the relationship between language and thought. In our era of fake news and corrupt media, this has only become even more pronounced: if you lie to a population and confuse them enough, you can control them. O’Brien introduces Winston to the work of the traitor to the Party, Emmanuel Goldstein, only to tell him later that Goldstein may not exist and his book was actually written by the Party.

Is this the lie, or was the book the lie? One of the most famous lines from the novel is Winston’s note to himself in his diary: ‘Freedom is the freedom to say that two plus two make four. If that is granted, all else follows.’

But later, O’Brien will force Winston to ‘admit’ that two plus two can make five. Orwell tells us, ‘The Party told you to reject the evidence of your eyes and ears.’

Or as Voltaire once wrote, ‘Truly, whoever is able to make you absurd is able to make you unjust.’ Forcing somebody to utter blatant falsehoods is a powerful psychological tool for totalitarian regimes because through doing so, they have chipped away at your moral and intellectual integrity.

Discover more from Interesting Literature

Subscribe to get the latest posts to your email.

Type your email…

4 thoughts on “A Summary and Analysis of George Orwell’s Nineteen Eighty-Four”

1984 is a novel which is great in spite of itself and has been lionised for the wrong reasons. The title of the novel is a simple anagram of 1948, the date when the novel was written, and was driven by Orwell’s paranoia about the 1945 Labour government in UK. Orwell, a public school man, had built a reputation for hiself in the nineteen thirties as a socialist writer, and had fought for socialism in the Spanish civil war. The Road To Wigan Pier is an excellent polemic attacking the way the UK government was handling the mass unemployment of the time, reducing workers to a state of near starvation. In Homage To Catalonia, Orwell describes his experiences fighting with a small Marxist militia against Franco’s fascists. It was in Spain that Orwell developed his lifelong hatred of Stalinism, observing that the Communist contingents were more interested in suppressing other left-wing factions than in defeating Franco. The 1945 Labour government ws Britain’s first democratically elected socialist governement. It successfully established the welfare state and the National Health Service in a country almost bankrupted by the war, and despite the fact that Truman in USA was demanding the punctual repayment of wartime loans. Instead of rejoicing, Orwell, by now terminally ill from tuberculosis, saw the necessary continuation of wartime austerity and rationing as a deliberate and unnecessary imposition. Consequently, the book is often used as propaganda against socialism. The virtues of the book are the warnings about the dangers of giving the state too much power, in the form of electronic surveillance, ehanced police powers, intrusive laws, and the insidious use of political propaganda to warp peoples’ thinking. All of this has come to pass in the West as well as the East, but because of the overtly anticommunist spin to Orwell’s novel, most people fail to get its important message..

As with other work here, another good review. I’m also fascinated that Orwell located the government as prime problem, whereas Huxley located the people as prime problem, two sides of the same coin.

• Pingback: Top 3 Historical Fiction Books – BookNook
• Pingback: 10 of the Best Books and Stories Set in the Future – Interesting Literature

Leave a Reply Cancel reply

Subscribe now to keep reading and get access to the full archive.

Continue reading

George Orwell's Anti-Utopian Reality in 1984 Novel

The overwhelming spread of military literature in the 20th century gave readers a great abundance of books to read on these topics. Some authors take both the pro and con sides of the military states and actions in discussing the political realities of their times. Among them, George Orwell wrote a novel that depicted the future that is relevant for all centuries and all political powers. The book 1984 (published in 1949, right after World War II) talks about a personality that has to survive under the pressures of an oppressive government.

About George Orwell

George Orwell, whose real name was Eric Arthur Blair, was an English novelist, essayist, journalist, and critic known for his keen observations on social injustice, totalitarianism, and democratic socialism. Born in India in 1903, Orwell spent much of his life in England and was deeply influenced by his experiences during the Spanish Civil War, where he fought against fascism.

George Orwell's life and career were marked by a commitment to truth-telling and a staunch opposition to propaganda and censorship. His experiences as a colonial police officer in Burma provided him with firsthand insight into the workings of the empire and the abuses of power. Orwell's disdain for authoritarianism extended to his critique of capitalism, evident in works such as Animal Farm, a satirical allegory of the Russian Revolution. Despite his socialist leanings, Orwell remained fiercely independent in his thinking, resisting ideological conformity and maintaining skepticism towards political movements of all stripes.

His writing style, characterized by clarity, precision, and mastery of language, continues to captivate readers and influence writers to this day. Through his literary legacy, 1984 stands as one of his most famous works, a dystopian masterpiece that continues to resonate with readers worldwide.

Throughout the whole story, Orwell depicts an invisible fight between the individual and the system. The book is pretty dark, heavy and depressing. Under enormous pressure, the protagonist of the story betrays his love, admits that 2+2 is 5 and glorifies his oppressors. He can’t afford an extra move, step, or look – Big Brother is watching him. The reader can get scared reading the book – but not reading it will leave all of us blind to the potential dangers of this world.

It would be mistaken to assume that 1984 makes a specific reference to one well-known social totalitarian state that no longer exists. The resistance to oppression was relevant before the USSR appeared; it is still relevant in many situations today and will still be relevant no matter how democratic and liberal our societies claim to be. That’s why 1984 was, is and will be the desk companion for many readers throughout the world.

Initially met with mixed reviews, the novel gradually gained widespread acclaim for its chilling portrayal of a dystopian future. Critics and scholars alike have praised Orwell's prescient vision of a totalitarian society where individual freedoms are systematically eroded, and truth becomes a malleable commodity. "1984" has been lauded for its incisive critique of surveillance, propaganda, and the abuse of power by authoritarian regimes.

Over the years, the book has become a cultural touchstone, inspiring countless adaptations, references in popular culture, and ongoing discussions about its relevance to contemporary political realities. Its themes of government overreach, thought control and resistance against oppression continue to resonate with readers worldwide, cementing "1984" as a timeless and indispensable work of literature.

What Is the Main Point in 1984?

The main point in George Orwell's "1984" revolves around the dangers of totalitarianism and the suppression of individual freedom. Set in a dystopian future where the ruling Party exerts complete control over every aspect of society, including language, history, and thought, the novel portrays a bleak world where truth is manipulated, dissent is punished, and surveillance is omnipresent.

Through the protagonist Winston Smith's journey of rebellion and disillusionment, Orwell underscores the importance of critical thinking, truth-seeking, and the inherent value of human autonomy. "1984" serves as a stark warning against the encroachment of oppressive governments on individual liberties, urging readers to remain vigilant against threats to freedom and to resist attempts to undermine the integrity of truth and independent thought.

Feeling like Your Essay's Playing Hard to Get?

No sweat! We've got the experts to help you out, whether it's penning, polishing, or sorting your ideas.

Main Characters and Roles of 1984

The characters of the book each serve very specific roles and purposes in the text, so let’s first briefly explore what the 1984 book is about. The book talks about a possible scenario for the development of the world. After several sanguinary wars and revolutions, the Earth was divided into 3 super states named Oceania, Eurasia, and Eastasia. Their alfa governments are in constant conflict with each other. Such never-ending conflicts are needed to distract the attention of the population from poor internal public management, terrible living conditions of the counties. More importantly, the existence of the conflict allows the government to fully control the inhabitants of the states.

Winston Smith Character Analysis

In one of such “superstates”, namely Oceania, lives the protagonist of the book. He is 39, he is thin and has a somewhat unhealthy look on his face. An employee of the Ministry of Truth, Winston Smith serves the government institution that works day and night to rewrite the past and destroy the facts that are unwanted by the government. Every day Winston changes the past with his own hands and makes it conform to the new standards devised by the ruling party.

In addition to changing the past, the Ministry of Truth also works tirelessly to promulgate the values and mantras of the county’s political elite. Seeing such truth tailoring and past elimination on a daily basis, Mr. Smith can’t help but wonder whether what is happening is right.

His soul grows a seed of suspicion and doubt and that induces him to start writing a diary. This diary is the only thing that hears what Winston thinks about his job, his life and his government, it marks the beginning of his protest.

The protagonist has to be very careful and do the writing in complete secrecy, hiding from other people and devices. As mentioned in Part 1 Chapter 1, his TV is not only a tool to feed him proper information, it also spies on him:

“The telescreen received and transmitted simultaneously. Any sound that Winston made, above the level of a very low whisper, would be picked up by it, moreover, so long as he remained within the field of vision which the metal plaque commanded, he could be seen as well as heard”.

Whatever he writes in his diary is a crime of through and qualifies for the death penalty.

Big Brother Character Analysis

Big Brother is the supreme ruler of Oceania. He has zero tolerance for individualism or diversity and absolutely no need for pluralism of opinion. He also has a network of Spies and tools set up in the country to make sure that every move of his citizens is observed, controlled and can be contained, if necessary. The Spies adore him and the Party:

Part 1, Chapter 2 “The songs, the processions, the banners, the hiking, the drilling with dummy rifles, the yelling of slogans, the worship of Big Brother — it was all a sort of glorious game to them.”

It’s impossible to do something privately in Oceania: all the houses are made of glass, all walls have surveillance and wiretapping, the Thought Police watches every move of every citizen. However, there is a difference in how Big Brother treats certain classes of its citizens. For example, for their love affair, Winston and Julia often choose secret places for dating, such as the countryside or other places where normally low-class labor workers hang out because the state doesn’t have that much security there. Low worker class is considered to have less tendency for thinking thus is treated as a lower-risk population.

Big Brother is an ultimate leader of Oceania, he is like a God and the ultimate goal is to please him. All the mistakes and loopholes of Big Brother or the Party are simply rewritten just like the newspapers. His pictures are everywhere, all the slogans are signed by his name. He is the only source of information, faith and worship in Oceania.

O'Brien Character Analysis

O’Brien is an undercover agent of the party. He secretly works for the Thought Police trying to find people who are thinking about rebellion. He is well-behaved, reserved, has a strong body. He deliberately pretends to oppose the party and Big Brother. His role is similar to that of Mephistopheles in Faust, he is the agent of the devil.

O’Brien is both a character and a concept in the book. He invades the dreams and provokes Smith to think that he doesn’t share Party ideas, he constantly pushes Smith to give birth to his unspoken internal conflict. Finally, when Smith and Julia are ready, he offers them to join the rebel movement. Later O’Brien will personally supervise the torture of his capturers, slowly killing any traces of personalities or thinking in them.

Emmanuel Goldstein Character Analysis

Emmanuel Goldstein was once a leader of the Party that brought it to power. He is now in exile and represents the only opposition available. He established an organization “Brotherhood” that is proclaimed by the Party to be the Enemy of the People. In fact, nobody knows for sure whether the organization really exists and what it does. Goldstein is an imaginary magnet for potential opposition, he serves the purpose of bringing all those who are against the Party under one roof to be destroyed then.

The Party spends a great deal of effort to publicly broadcast the hate clips about Goldstein and the Brotherhood just to give a bait for those who are seeking allies to create a rebellion.

Tom Parsons Character Analysis

Tom Parsons and his wife Mrs. Parsons live next door to Winston. Tom is a complete opposite of Smith, he follows the Party blindly and never doubts Oceania for a second. He is devoted to the war against other states and will do whatever he can to contribute to Oceania’s victory.

Ironically, he brought up a daughter who is just as fierce and loyal to Oceania as her parents are. One day she betrays her father by reporting to the Thought Police that Parsons spoke badly of Big Brother in his sleep. To aggravate the irony even more, Orwell makes Tome immensely proud of his daughter for “doing the right thing”.

Our research paper writer service is ready for any requests. Push the button below to contact a professional writer.

Julia Character Analysis

Julia is another protagonist of 1984. She is 26, she also works for the Ministry of Truth in the Fiction Department. She writes novels depicting the greatness of her country and its ruler. She is quite experienced sexually and is known to seduce Party members. She is instinctive, not very logical, irrational, with lots of untamed desire and energy. She is courageous and much more adventurous than her lover Smith. In fact, she is the one who tells about her feelings to Winston and takes him outside of town.

It’s difficult to elaborate on the nature of Julia’s and Winston’s relationship since they are the only creatures with a soul portrayed in this book. So it makes sense that they found each other and grew fond of each other. Would they have felt just as fond of each other if there were other options available – who knows? But the main point Orwell makes is that in such an authoritarian government as Oceania, finding people who think and have their own opinion is an extremely rare thing.

Julia’s sexual and emotional freedom is her way to protest against the strict order of her country. She wants to put her energy into love, emotions, memories and enjoyment, not for the glorification of Big Brother and Oceania. And it only makes the reader even more upset when in the end she breaks under the tortures of O’Brien and says in Part 3 Chapter 6:

“You think there's no other way of saving yourself, and you're quite ready to save yourself that way. You want it to happen to the other person. You don't give a damn what they suffer. All you care about is yourself”.

Mr. Charrington Character Analysis

Mr. Charrington is the owner of a thrift shop in a parole district. Proles are the majority of Oceania population who are not part of the Inner Party (those who rule) or Outer Party (those who serve the rulers) and are deemed incapable of thinking or posing a threat to the government. However, in Part 1 Chapter 7 Winston expressed his opinion in the diary that proles might rebel one day and take the Party down:

“If there is hope, it lies in the proles”.

Winston buys his diary from Mr. Charrington and that marks the beginning of Winston’s journey into critical thinking and rebellion. Later, Winston will rent a bedroom upstairs above the shop to meet with Julia there.

Winston trusts Mr. Charrington because he holds on to the past (second-hand items) and thus keeps the past intact when Oceania is doing everything it can to change or destroy the past. At some point, Winston even thinks that Mr. Charrington is a member of the Brotherhood. But as it turns out, he is an informant of the Police and spies on everything Winston and Julia do in his shop.

Do you know anything about the AMA paper format ? We recommend reading the article from our authors.

1984 Full Summary

After the Second World War, the civil war broke down in Great Britain, which lead to it being occupied by a new superstate – Oceania. The citizens of Oceania live under the rule of an ideology of one Party. The ruler and impersonification of that Party is a leader called Big Brother.

The Party is divided into Inner Party (the 2% of the ruling population), Outer Party (the 13% who implement their policies) and the others, who are called proles and don’t have any opinion or importance whatsoever. But not all members of the Outer Party are in unanimous agreement with the Party ideology. Winston Smith works for the Ministry of Truth and is starting to question the Party’s right to rule and tell him what to do. But he understands that there’s nobody with whom he can share his concerns. So he shares his thoughts in a diary, which is also quite a dangerous thing to do.

One day Smith notices that his colleague Julia is paying a lot of attention to him. At first, he is afraid that she busted him and will give him up for the Thought Police. But after some time he finds a love note from her. They start a secret relationship that is prohibited by the government. They hide and dream about a revolution. Smith believes that their relationship will not end well – such encounters between men and women are strictly prohibited in Oceania.

Eventually, they meet a representative of a real revolutionary movement, O’Brien, who gives them a book on the philosophy of the upcoming rebellion. While reading the book in the room they rented for dating, the couple is busted by the Through Police – the so-called revolution movement representative was nothing but a set-up of Big Brother to find and eliminate potential rebels.

The government imprisons Julia and Winston and tortures them cruelly. They break under the tortures and betray each other. In the end, both Winston and his ex-beloved Julia praise the majesty and powerfulness of Big Brother and sincerely believe that their country is doing great. The Through Police manages to “cure” Winston from his revolutionary thoughts. At first, Smith thinks that he gave up Julia and his freedom just to evade the torture, but once he is released, he realizes that he is now the right man who sincerely believes in Big Brother and the Party.

EssayPro is a writing paper service , so we can assist you in any kind of essays or papers.

1984 Essay Sample

Dive into a literary exploration with our essay sample, where '1984' is examined under the microscope of expert analysis.

1984 Theme 1: War. The author wrote his dystopian classic in 1948 and he simply changes the last two digits of the year when naming his book. The first theme that is present in the text is the war – 1948 is the time after one of the biggest tragedies in human history, Second World War, and the time when the world watched in terror the emergence of two huge military powers – USA and USSR. Despite the victory and defeat of the fascist movement, people, tired of the loss and tragedy the WW2 brought about, felt helpless when it came to the conception of potential World War Three. The danger was in the air, the fatigue was in the minds, the fear was in the nightmares lived by almost everybody around the world. 1984 was just one of the many military literature pieces heavily exploring one of the possible scenarios that were about to happen.

In 1984 there are three states — two of which are allied, while the third is an enemy. The alliances change regularly and yesterday’s ally can turn into an enemy tomorrow. The war and conflict give Oceania a powerful excuse to disregard the shortages of food, ever-present surveillance and other social problems. The war is a guarantee of internal order in Oceania – how can a loyal citizen undermine his own country when they are at war with an external enemy?

1984 theme 2: Control. Dictatorship and the right of any institution or any given personality to exercise control over people was a hot topic for discussion towards the end of the 20th century. The thing is that there are people who don’t like making decisions because with decisions comes responsibility. So they welcome others to make decisions for them and society accepts it as their right to use predefined solutions. But step by step such willingness to let others make your choices can turn into a dangerous overcontrolling net. Oceania didn’t appear in one day, some processes led to it being like we know it. In 1984 Orwell elaborates what consequences can the war between authoritarian states have and how easy it is to turn to tyranny “for the greater good of the society”.

The citizens of Oceania are in the absolute unity with their state: if they are following the state, they have nothing to worry about, nothing to hide, nothing to think about. They are the state, and the state is at war – so when Oceania wins the war, they will win as well. The control chain is eternal.

1984 theme 3: Mind Control through Newspeak language. The overwhelming control over social life was enhanced through another theme heavily explored by Orwell – the creation of a new language for Oceania called Newspeak. The new English Socialism ideology developed by the ruling party was imposed through the invention of its own language, where each word and grammatical rule were carefully handpicked. When the events in the book took place, the new language was in the process of being introduced: it appeared in the newspapers and party members wouldn’t miss an opportunity to insert a phrase or two in their speeches. The Newspeak was supposed to have completely replaced the Oldspeak (regular English language known and spoken today and in 1980s) by 2050. That would mean yet another victory of Oceania over people’s minds and freedoms.

1984 theme 4: New and improved truth. To keep the society in place and make sure the country is not disturbed and remains focused on the war with another state, the employees of the Ministry of Truth change the news. Every day they rewrite the newspapers of yesterday, backdate them and put them back into circulation.

The altered truth concept is also revealed in the fact that Winston is not actually that good of a character. He wants to be able to think and to love, but the truth is that he is also a wicked personality: he used to steal food from his mother and sisters, he ran away from home. And the readers aren’t sure whether he regrets doing it or not.

Perhaps you need help writing a student paper, such as an SAT essay ? You can get paper writing help from EssayPro.

Symbolism in 1984

The Memory Hole

Winston’s job was about changing the news so that it matched the reality that Oceania wanted its citizens to see. In his office there were three holes in the wall: for notes on changes that had to be made, for newspapers that had to be edited and for recycling of all the materials. They were called “memory holes” as symbols of ways to destroy and alter memories of thousands of people. Memory holes are also symbols for distorted communication channels Oceania used to brainwash its citizens.

Big Brother

There was one recognizable face that appeared on numerous propaganda materials (posters, TV clips, newspapers and etc.). These materials persuaded citizens how great Oceania was and also delivered a message that “he is watching” everybody at all times. It’s a message of hope (the country will be great one day) and desperation (you are watched 24/7). Big Brother is a symbol of Oceania’s national agenda, he is an idol, a person who gained enormous power not due to his leadership potential, but because of Oceania’s inhumate treatment of its citizens.

Winston had to admit to this famous calculation when he was tortured by the Though Police. This is the symbol of a vivid false statement that is accepted socially in the society governed by a totalitarian ideology.

Winston's Varicose Ulcer

The medical condition that bothers Winston represents his oppressed feelings and desires. It is an external expression of his internal pains. From one point of view, varicose ulcer is a symbol of Smith sexual desire that is prohibited to exhibit in Oceania. On another hand, it’s a mark of Winston’s dissatisfaction with what is going on around him, it’s a visible physical repercussion of living under total control.

The Red-armed Singing Prole Woman

The woman from a lower worker class (prole) is a symbol of potential rebellion. Winston believed that proles would rebel one day and that the hope for Oceania to regain its civic freedoms lies with proles. Her female capacity to give birth is a symbol that a thought can be born within proles’ minds and new generations can see the world without total control of Big Brother.

1984 is a book that will live forever. It will resonate with readers from different countries, backgrounds, and political views. It is an instruction for government managers on how to compel obedience from its citizens. It’s also a vivid demonstration for citizens how the government can make them do whatever. It’s a scary but real story, cruel but eye-opening, it changes the way we treat our fundamental freedom rights. This book helps us appreciate what we have – the possibility to choose friends, love the people we find attractive, do what we like doing, think, speak, and make decisions in our lives.

Sometimes when you have a task to write custom essays according to the book plot, and it's complicated for you to understand the idea of the book, you may need help. It can be writing, editing, proofreading, etc. So contact us to hire a persuasive essay writer and we will help you asap.

Feeling lost when trying to grasp the essence of a book for your essay? No need to fret. Whether it's writing, editing, or just getting your thoughts in order, we're here to lend a hand. Reach out to us to hire a dedicated persuasive essay writer who can guide you through understanding and expressing your ideas clearly. We're here to help, so don't hesitate to get in touch for custom essays !

Lost in the Jungle of Essay Writing?

Don't wander alone! We're here to navigate the twists and turns of writing, editing, and brainstorming with you.

Where Does 1984 Take Place?

What is the key idea of 1984.

is an expert in nursing and healthcare, with a strong background in history, law, and literature. Holding advanced degrees in nursing and public health, his analytical approach and comprehensive knowledge help students navigate complex topics. On EssayPro blog, Adam provides insightful articles on everything from historical analysis to the intricacies of healthcare policies. In his downtime, he enjoys historical documentaries and volunteering at local clinics.

Related Articles

1984 Themes – Meaning and Main Ideas

Home » Literature Explained – Literary Synopses and Book Summaries » 1984 Book » 1984 Themes – Meaning and Main Ideas

Main Theme of 1984 – Introduction

The novel takes place in a futuristic and dystopian version of London, UK. The citizens of this nation, Oceania, are ruled by Big Brother and The Party. They are under constant surveillance and the information that they receive is controlled by The Party before it reaches any citizens. The novel was written in 1949 but the exact year of the story is unknown. Even the main character, Winston, is unsure of the exact date anymore because The Party keeps its citizens uninformed and he lost track. We know that it is “the future” because of all of the technology and the title leads us to guess it may be in the year 1984, which shows Orwell’s intentional message that a government takeover with advanced technologies could be more imminent than anyone would want to believe.

There are several very strong themes in this short novel, and a couple of motifs that back those themes up and support the overall message Orwell intended to create. Doublethink is a motif in the novel—it occurs when The Party suddenly changes the information that they’ve been giving the citizens. The citizens agree to just go along with the changes and are able to believe whatever they need too, even if it is all directly contradictory. For example, when a speech is being given, the orator randomly changes which nation he refers to as their enemy. The people believe it right away and feel bad that they made the wrong signs to bring to the speech. Another motif is the decay of the city as a result of the violent revolution that occurred some years prior. The city is in a state of decay, but The Party ignores this, mismanaging a city of the size completely. This leaves the proles (lower class citizens) largely unmonitored, which is an oversight on the part of the government because it poses the potential for revolution.

Main Themes in 1984

Here’s a list of major themes in 1984.

• Totalitarianism
• State control of expression
• Control over information

Individual Identity

The inherent destruction in totalitarianism.

Psychological Manipulation Through Technology

State Control Over Expression

Since The Party is always watching, they also control how citizens use their bodies. They cannot have sex outside of procreation, and even a misgiving facial twitch could lead to an arrest and subsequent torture to break that individual into submission. The Party also requires daily exercises from all citizens, and they will be yelled at through their telescreens if they do not exercise hard enough. When people turn to anti-Party activities, they will be tortured by officials until they relent and show full brainwashed support for The Party.

Control Over Information

The Party has decided to control all information, being very careful what kinds of history the citizens are able to access. They develop Newspeak, which is a modified form of English that eliminates any words that could threaten The Party’s control over its people. People’s memories become fuzzy, they lose track of the year, and eventually they just comply because they don’t know any better.

The novel centers on Winston’s various acts of resistance that start small but then become bolder and bolder until he is finally arrested and tortured for it. He dreams of revolution, imagining that the proles will be the key to overthrowing The Party and giving future generations freedom. He finds inspiration in items that remind him of the past, which he can barely remember. He starts up a love affair with the beautiful Julia. All of these things lead Winston to seek out an anti-Party movement. Ultimately, though, he is arrested by double agents and this desire to resist is tortured out of him. The Party does not treat any opposition lightly, making sure to use every method they can possibly find to brainwash and remove desire for resistance in their citizens.

This website uses cookies

We place some essential cookies on your device to make this website work. We'd like to use additional cookies to remember your settings and understand how you use our services. This information will help us make improvements to the website.

Catalogue description Record Type: Memorandum Former Reference: C (82) 18 Title: Falkland...

Ordering and viewing options.

• Download format PDF
• Approximate size 0.5 MB
• Add to basket

Order up to 10 items per basket, and up to 100 in a 30 day period.

• Available to view free at The National Archives

Image viewer

This record is available to view with our image viewer. Show images

Preview a low-resolution image of this record.

Have you found an error with this catalogue description? Let us know

Your details

If you provide contact details we may contact you if we need more information.

Help with your research

• How to use this catalogue
• Need more context? View the catalogue description for CAB 129

Sign me up to the mailing list

Subscribe now for regular news, updates and priority booking for events

Find out more

• Jobs and careers
• Friends of The National Archives
• Website A-Z index
• Accessibility
• Podcasts and videos
• Image library
• UK Government Web Archive
• Legislation.gov.uk Opens a new window
• Find case law
• The Gazette Opens a new window
• Terms of use
• Privacy policy
• Freedom of Information
• Transparency

All content is available under the Open Government Licence v3.0 , except where otherwise stated

Help | Advanced Search

Computer Science > Computation and Language

Title: chameleon: mixed-modal early-fusion foundation models.

Abstract: We present Chameleon, a family of early-fusion token-based mixed-modal models capable of understanding and generating images and text in any arbitrary sequence. We outline a stable training approach from inception, an alignment recipe, and an architectural parameterization tailored for the early-fusion, token-based, mixed-modal setting. The models are evaluated on a comprehensive range of tasks, including visual question answering, image captioning, text generation, image generation, and long-form mixed modal generation. Chameleon demonstrates broad and general capabilities, including state-of-the-art performance in image captioning tasks, outperforms Llama-2 in text-only tasks while being competitive with models such as Mixtral 8x7B and Gemini-Pro, and performs non-trivial image generation, all in a single model. It also matches or exceeds the performance of much larger models, including Gemini Pro and GPT-4V, according to human judgments on a new long-form mixed-modal generation evaluation, where either the prompt or outputs contain mixed sequences of both images and text. Chameleon marks a significant step forward in a unified modeling of full multimodal documents.

Submission history

Access paper:.

• HTML (experimental)
• Other Formats

References & Citations

• Google Scholar
• Semantic Scholar

BibTeX formatted citation

Bibliographic and Citation Tools

Code, data and media associated with this article, recommenders and search tools.

• Institution

arXivLabs: experimental projects with community collaborators

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs .

Main Navigation

• Contact NeurIPS
• Code of Ethics
• Code of Conduct
• Create Profile
• Journal To Conference Track
• Diversity & Inclusion
• Proceedings
• Future Meetings
• Exhibitor Information
• Privacy Policy
• Visualization
• SurDis: A Surface Discontinuity Dataset for Wearable Technology to Assist Blind Navigation in Urban Environments
• MOMA-LRG: Language-Refined Graphs for Multi-Object Multi-Actor Activity Parsing
• VLMbench: A Compositional Benchmark for Vision-and-Language Manipulation
• AnimeRun: 2D Animation Visual Correspondence from Open Source 3D Movies
• Pythae: Unifying Generative Autoencoders in Python - A Benchmarking Use Case
• Ambiguous Images With Human Judgments for Robust Visual Event Classification
• Towards Better Evaluation for Dynamic Link Prediction
• pyKT: A Python Library to Benchmark Deep Learning based Knowledge Tracing Models
• EgoTaskQA: Understanding Human Tasks in Egocentric Videos
• Finding Naturally Occurring Physical Backdoors in Image Datasets
• Characteristics of Harmful Text: Towards Rigorous Benchmarking of Language Models
• GLOBEM Dataset: Multi-Year Datasets for Longitudinal Human Behavior Modeling Generalization
• K-Radar: 4D Radar Object Detection for Autonomous Driving in Various Weather Conditions
• PulseImpute: A Novel Benchmark Task for Pulsative Physiological Signal Imputation
• Touch and Go: Learning from Human-Collected Vision and Touch
• How Transferable are Video Representations Based on Synthetic Data?
• Nocturne: a scalable driving benchmark for bringing multi-agent learning one step closer to the real world
• How Would The Viewer Feel? Estimating Wellbeing From Video Scenarios
• The BigScience ROOTS Corpus: A 1.6TB Composite Multilingual Dataset
• A Comprehensive Study on Large-Scale Graph Training: Benchmarking and Rethinking
• SMPL: Simulated Industrial Manufacturing and Process Control Learning Environments
• Hard ImageNet: Segmentations for Objects with Strong Spurious Cues
• SCAMPS: Synthetics for Camera Measurement of Physiological Signals
• OpenOOD: Benchmarking Generalized Out-of-Distribution Detection
• A Large Scale Search Dataset for Unbiased Learning to Rank
• FinRL-Meta: Market Environments and Benchmarks for Data-Driven Financial Reinforcement Learning
• Tenrec: A Large-scale Multipurpose Benchmark Dataset for Recommender Systems
• MineDojo: Building Open-Ended Embodied Agents with Internet-Scale Knowledge
• Robustness Analysis of Video-Language Models Against Visual and Language Perturbations
• ComMU: Dataset for Combinatorial Music Generation
• BOND: Benchmarking Unsupervised Outlier Node Detection on Static Attributed Graphs
• TAP-Vid: A Benchmark for Tracking Any Point in a Video
• DABS 2.0: Improved Datasets and Algorithms for Universal Self-Supervision
• Communicating Natural Programs to Humans and Machines
• Benchmarking Heterogeneous Treatment Effect Models through the Lens of Interpretability
• SkinCon: A skin disease dataset densely annotated by domain experts for fine-grained debugging and analysis
• PROSPECT: Labeled Tandem Mass Spectrometry Dataset for Machine Learning in Proteomics
• FACT: Learning Governing Abstractions Behind Integer Sequences
• Turning the Tables: Biased, Imbalanced, Dynamic Tabular Datasets for ML Evaluation
• PeRFception: Perception using Radiance Fields
• A Greek Parliament Proceedings Dataset for Computational Linguistics and Political Analysis
• Addressing Resource Scarcity across Sign Languages with Multilingual Pretraining and Unified-Vocabulary Datasets
• TGEA 2.0: A Large-Scale Diagnostically Annotated Dataset with Benchmark Tasks for Text Generation of Pretrained Language Models
• BackdoorBench: A Comprehensive Benchmark of Backdoor Learning
• 3DOS: Towards 3D Open Set Learning - Benchmarking and Understanding Semantic Novelty Detection on Point Clouds
• Flare7K: A Phenomenological Nighttime Flare Removal Dataset
• DC-BENCH: Dataset Condensation Benchmark
• Change Event Dataset for Discovery from Spatio-temporal Remote Sensing Imagery
• CLiMB: A Continual Learning Benchmark for Vision-and-Language Tasks
• Understanding Aesthetics with Language: A Photo Critique Dataset for Aesthetic Assessment
• IKEA-Manual: Seeing Shape Assembly Step by Step
• FLamby: Datasets and Benchmarks for Cross-Silo Federated Learning in Realistic Healthcare Settings
• Ontologue: Declarative Benchmark Construction for Ontological Multi-Label Classification
• MBW: Multi-view Bootstrapping in the Wild
• Enabling Detailed Action Recognition Evaluation Through Video Dataset Augmentation
• Sample Efficiency Matters: A Benchmark for Practical Molecular Optimization
• NAS-Bench-Suite-Zero: Accelerating Research on Zero Cost Proxies
• PEER: A Comprehensive and Multi-Task Benchmark for Protein Sequence Understanding
• FlyView: a bio-informed optical flow truth dataset for visual navigation using panoramic stereo vision
• Chartalist: Labeled Graph Datasets for UTXO and Account-based Blockchains
• A Multi-Task Benchmark for Korean Legal Language Understanding and Judgement Prediction
• APT-36K: A Large-scale Benchmark for Animal Pose Estimation and Tracking
• OccGen: Selection of Real-world Multilingual Parallel Data Balanced in Gender within Occupations
• GriddlyJS: A Web IDE for Reinforcement Learning
• Model Zoos: A Dataset of Diverse Populations of Neural Network Models
• Meta-Album: Multi-domain Meta-Dataset for Few-Shot Image Classification
• Beyond Real-world Benchmark Datasets: An Empirical Study of Node Classification with GNNs
• The Surprising Effectiveness of PPO in Cooperative Multi-Agent Games
• Breaking Bad: A Dataset for Geometric Fracture and Reassembly
• USB: A Unified Semi-supervised Learning Benchmark for Classification
• NAS-Bench-Graph: Benchmarking Graph Neural Architecture Search
• WinoGAViL: Gamified Association Benchmark to Challenge Vision-and-Language Models
• BigBio: A Framework for Data-Centric Biomedical Natural Language Processing
• MATE: Benchmarking Multi-Agent Reinforcement Learning in Distributed Target Coverage Control
• CLEVRER-Humans: Describing Physical and Causal Events the Human Way
• HandMeThat: Human-Robot Communication in Physical and Social Environments
• OpenSRH: optimizing brain tumor surgery using intraoperative stimulated Raman histology
• Pile of Law: Learning Responsible Data Filtering from the Law and a 256GB Open-Source Legal Dataset
• BLOX: Macro Neural Architecture Search Benchmark and Algorithms
• CEDe: A collection of expert-curated datasets with atom-level entity annotations for Optical Chemical Structure Recognition
• M4Singer: A Multi-Style, Multi-Singer and Musical Score Provided Mandarin Singing Corpus
• LAION-5B: An open large-scale dataset for training next generation image-text models
• MSDS: A Large-Scale Chinese Signature and Token Digit String Dataset for Handwriting Verification
• Benchmarking and Analyzing 3D Human Pose and Shape Estimation Beyond Algorithms
• ViSioNS: Visual Search in Natural Scenes Benchmark
• mRI: Multi-modal 3D Human Pose Estimation Dataset using mmWave, RGB-D, and Inertial Sensors
• ActionSense: A Multimodal Dataset and Recording Framework for Human Activities Using Wearable Sensors in a Kitchen Environment
• The Dollar Street Dataset: Images Representing the Geographic and Socioeconomic Diversity of the World
• FETA: Towards Specializing Foundational Models for Expert Task Applications
• AnoShift: A Distribution Shift Benchmark for Unsupervised Anomaly Detection
• Evaluating Out-of-Distribution Performance on Document Image Classifiers
• Open High-Resolution Satellite Imagery: The WorldStrat Dataset – With Application to Super-Resolution
• OLIVES Dataset: Ophthalmic Labels for Investigating Visual Eye Semantics
• A Benchmark for Compositional Visual Reasoning
• xView3-SAR: Detecting Dark Fishing Activity Using Synthetic Aperture Radar Imagery
• Learning Long-Term Crop Management Strategies with CyclesGym
• ETAB: A Benchmark Suite for Visual Representation Learning in Echocardiography
• EHRSQL: A Practical Text-to-SQL Benchmark for Electronic Health Records
• GOOD: A Graph Out-of-Distribution Benchmark
• Is one annotation enough? - A data-centric image classification benchmark for noisy and ambiguous label estimation
• MTNeuro: A Benchmark for Evaluating Representations of Brain Structure Across Multiple Levels of Abstraction
• CAESAR: An Embodied Simulator for Generating Multimodal Referring Expression Datasets
• JAHS-Bench-201: A Foundation For Research On Joint Architecture And Hyperparameter Search
• A Dataset for Efforts Towards Achieving the Sustainable Development Goal of Safe Working Environments
• Forecasting Future World Events With Neural Networks
• TwiBot-22: Towards Graph-Based Twitter Bot Detection
• Avalon: A Benchmark for RL Generalization Using Procedurally Generated Worlds
• Long Range Graph Benchmark
• Geoclidean: Few-Shot Generalization in Euclidean Geometry
• CARLANE: A Lane Detection Benchmark for Unsupervised Domain Adaptation from Simulation to multiple Real-World Domains
• EnvPool: A Highly Parallel Reinforcement Learning Environment Execution Engine
• How Well Do Unsupervised Learning Algorithms Model Human Real-time and Life-long Learning?
• OpenFilter: A Framework to Democratize Research Access to Social Media AR Filters
• Why do tree-based models still outperform deep learning on typical tabular data?
• Multi-LexSum: Real-world Summaries of Civil Rights Lawsuits at Multiple Granularities
• Wukong: A 100 Million Large-scale Chinese Cross-modal Pre-training Benchmark
• Robustness Disparities in Face Detection
• AMOS: A Large-Scale Abdominal Multi-Organ Benchmark for Versatile Medical Image Segmentation
• TaiSu: A 166M Large-scale High-Quality Dataset for Chinese Vision-Language Pre-training
• Wild-Time: A Benchmark of in-the-Wild Distribution Shift over Time
• PDEBench: An Extensive Benchmark for Scientific Machine Learning
• LIPS - Learning Industrial Physical Simulation benchmark suite
• Towards Video Text Visual Question Answering: Benchmark and Baseline
• SoundSpaces 2.0: A Simulation Platform for Visual-Acoustic Learning
• NeoRL: A Near Real-World Benchmark for Offline Reinforcement Learning
• NAS-Bench-360: Benchmarking Neural Architecture Search on Diverse Tasks
• OpenFWI: Large-scale Multi-structural Benchmark Datasets for Full Waveform Inversion
• METS-CoV: A Dataset of Medical Entity and Targeted Sentiment on COVID-19 Related Tweets
• DGraph: A Large-Scale Financial Dataset for Graph Anomaly Detection
• ConfLab: A Data Collection Concept, Dataset, and Benchmark for Machine Analysis of Free-Standing Social Interactions in the Wild
• HAPI: A Large-scale Longitudinal Dataset of Commercial ML API Predictions
• TempEL: Linking Dynamically Evolving and Newly Emerging Entities
• ELEVATER: A Benchmark and Toolkit for Evaluating Language-Augmented Visual Models
• A Survey and Datasheet Repository of Publicly Available US Criminal Justice Datasets
• Myriad: a real-world testbed to bridge trajectory optimization and deep learning
• TweetNERD - End to End Entity Linking Benchmark for Tweets
• AutoWS-Bench-101: Benchmarking Automated Weak Supervision with 100 Labels
• SafeBench: A Benchmarking Platform for Safety Evaluation of Autonomous Vehicles
• This is the way: designing and compiling LEPISZCZE, a comprehensive NLP benchmark for Polish
• A Unified Evaluation of Textual Backdoor Learning: Frameworks and Benchmarks
• Kantorovich Strikes Back! Wasserstein GANs are not Optimal Transport?
• DART: Articulated Hand Model with Diverse Accessories and Rich Textures
• Active-Passive SimStereo - Benchmarking the Cross-Generalization Capabilities of Deep Learning-based Stereo Methods
• CGLB: Benchmark Tasks for Continual Graph Learning
• ADBench: Anomaly Detection Benchmark
• A new dataset for multilingual keyphrase generation
• Unravelling the Performance of Physics-informed Graph Neural Networks for Dynamical Systems
• DDXPlus: A New Dataset For Automatic Medical Diagnosis
• Video compression dataset and benchmark of learning-based video-quality metrics
• Towards Human-Level Bimanual Dexterous Manipulation with Reinforcement Learning
• MVP-N: A Dataset and Benchmark for Real-World Multi-View Object Classification
• pFL-Bench: A Comprehensive Benchmark for Personalized Federated Learning
• Dungeons and Data: A Large-Scale NetHack Dataset
• OpenXAI: Towards a Transparent Evaluation of Model Explanations
• Honor of Kings Arena: an Environment for Generalization in Competitive Reinforcement Learning
• ENS-10: A Dataset For Post-Processing Ensemble Weather Forecasts
• AirfRANS: High Fidelity Computational Fluid Dynamics Dataset for Approximating Reynolds-Averaged Navier–Stokes Solutions
• EPIC-KITCHENS VISOR Benchmark: VIdeo Segmentations and Object Relations
• Multilingual Abusive Comment Detection at Scale for Indic Languages
• MoCapAct: A Multi-Task Dataset for Simulated Humanoid Control
• FLAIR: Federated Learning Annotated Image Repository
• StrokeRehab: A Benchmark Dataset for Sub-second Action Identification
• Training Uncertainty-Aware Classifiers with Conformalized Deep Learning
• Optimizing Relevance Maps of Vision Transformers Improves Robustness
• Quantum Speedups of Optimizing Approximately Convex Functions with Applications to Logarithmic Regret Stochastic Convex Bandits
• Low-rank lottery tickets: finding efficient low-rank neural networks via matrix differential equations
• Towards Improving Faithfulness in Abstractive Summarization
• SIREN: Shaping Representations for Detecting Out-of-Distribution Objects
• Implicit Neural Representations with Levels-of-Experts
• Uplifting Bandits
• Infinite-Fidelity Coregionalization for Physical Simulation
• RSA: Reducing Semantic Shift from Aggressive Augmentations for Self-supervised Learning
• On the Effective Number of Linear Regions in Shallow Univariate ReLU Networks: Convergence Guarantees and Implicit Bias
• On Infinite Separations Between Simple and Optimal Mechanisms
• TANKBind: Trigonometry-Aware Neural NetworKs for Drug-Protein Binding Structure Prediction
• Boosting the Transferability of Adversarial Attacks with Reverse Adversarial Perturbation
• Automatic differentiation of nonsmooth iterative algorithms
• Efficient coding, channel capacity, and the emergence of retinal mosaics
• Decentralized Local Stochastic Extra-Gradient for Variational Inequalities
• Toward Equation of Motion for Deep Neural Networks: Continuous-time Gradient Descent and Discretization Error Analysis
• Synthetic Model Combination: An Instance-wise Approach to Unsupervised Ensemble Learning
• Self-supervised Heterogeneous Graph Pre-training Based on Structural Clustering
• Tiered Reinforcement Learning: Pessimism in the Face of Uncertainty and Constant Regret
• Counterfactual Neural Temporal Point Process for Estimating Causal Influence of Misinformation on Social Media
• Globally Gated Deep Linear Networks
• Graph Scattering beyond Wavelet Shackles
• Aligning individual brains with fused unbalanced Gromov Wasserstein
• SoftPatch: Unsupervised Anomaly Detection with Noisy Data
• Kernel Interpolation with Sparse Grids
• Ask4Help: Learning to Leverage an Expert for Embodied Tasks
• TUSK: Task-Agnostic Unsupervised Keypoints
• Concept Activation Regions: A Generalized Framework For Concept-Based Explanations
• Matrix Multiplicative Weights Updates in Quantum Zero-Sum Games: Conservation Laws & Recurrence
• Posted Pricing and Dynamic Prior-independent Mechanisms with Value Maximizers
• Training stochastic stabilized supralinear networks by dynamics-neutral growth
• Chefs' Random Tables: Non-Trigonometric Random Features
• NeuForm: Adaptive Overfitting for Neural Shape Editing
• STaR: Bootstrapping Reasoning With Reasoning
• A Causal Analysis of Harm
• Network change point localisation under local differential privacy
• DISCO: Adversarial Defense with Local Implicit Functions
• Does GNN Pretraining Help Molecular Representation?
• FedAvg with Fine Tuning: Local Updates Lead to Representation Learning
• GET3D: A Generative Model of High Quality 3D Textured Shapes Learned from Images
• Re-Analyze Gauss: Bounds for Private Matrix Approximation via Dyson Brownian Motion
• Locating and Editing Factual Associations in GPT
• Faster Linear Algebra for Distance Matrices
• Causal Inference with Non-IID Data using Linear Graphical Models
• Extra-Newton: A First Approach to Noise-Adaptive Accelerated Second-Order Methods
• ALMA: Hierarchical Learning for Composite Multi-Agent Tasks
• Diversified Recommendations for Agents with Adaptive Preferences
• Optimizing Data Collection for Machine Learning
• VeriDark: A Large-Scale Benchmark for Authorship Verification on the Dark Web
• CoNT: Contrastive Neural Text Generation
• Trajectory-guided Control Prediction for End-to-end Autonomous Driving: A Simple yet Strong Baseline
• Zeroth-Order Negative Curvature Finding: Escaping Saddle Points without Gradients
• Towards Practical Control of Singular Values of Convolutional Layers
• Riemannian Neural SDE: Learning Stochastic Representations on Manifolds
• Pragmatically Learning from Pedagogical Demonstrations in Multi-Goal Environments
• A Contrastive Framework for Neural Text Generation
• AnimeSR: Learning Real-World Super-Resolution Models for Animation Videos
• Two-Stream Network for Sign Language Recognition and Translation
• Multivariate Time-Series Forecasting with Temporal Polynomial Graph Neural Networks
• Towards Out-of-Distribution Sequential Event Prediction: A Causal Treatment
• Roadblocks for Temporarily Disabling Shortcuts and Learning New Knowledge
• Learning Bipartite Graphs: Heavy Tails and Multiple Components
• Vision Transformers provably learn spatial structure
• Shape, Light, and Material Decomposition from Images using Monte Carlo Rendering and Denoising
• Teach Less, Learn More: On the Undistillable Classes in Knowledge Distillation
• Hand-Object Interaction Image Generation
• Feature Learning in $L_2$-regularized DNNs: Attraction/Repulsion and Sparsity
• Bridging the Gap: Unifying the Training and Evaluation of Neural Network Binary Classifiers
• On the Discrimination Risk of Mean Aggregation Feature Imputation in Graphs
• Efficient and Modular Implicit Differentiation
• NeuroSchedule: A Novel Effective GNN-based Scheduling Method for High-level Synthesis
• Recursive Reinforcement Learning
• Making Sense of Dependence: Efficient Black-box Explanations Using Dependence Measure
• Distribution-Informed Neural Networks for Domain Adaptation Regression
• On the Interpretability of Regularisation for Neural Networks Through Model Gradient Similarity
• Exploiting Semantic Relations for Glass Surface Detection
• Doubly-Asynchronous Value Iteration: Making Value Iteration Asynchronous in Actions
• Function Classes for Identifiable Nonlinear Independent Component Analysis
• GMMSeg: Gaussian Mixture based Generative Semantic Segmentation Models
• Recovering Private Text in Federated Learning of Language Models
• Contrastive Language-Image Pre-Training with Knowledge Graphs
• Fast Mixing of Stochastic Gradient Descent with Normalization and Weight Decay
• Disentangling the Predictive Variance of Deep Ensembles through the Neural Tangent Kernel
• Quantile Constrained Reinforcement Learning: A Reinforcement Learning Framework Constraining Outage Probability
• Neural Payoff Machines: Predicting Fair and Stable Payoff Allocations Among Team Members
• Diversity vs. Recognizability: Human-like generalization in one-shot generative models
• SketchBoost: Fast Gradient Boosted Decision Tree for Multioutput Problems
• DeVRF: Fast Deformable Voxel Radiance Fields for Dynamic Scenes
• Emergence of Hierarchical Layers in a Single Sheet of Self-Organizing Spiking Neurons
• Adapting to Online Label Shift with Provable Guarantees
• Offline Multi-Agent Reinforcement Learning with Knowledge Distillation
• Visual correspondence-based explanations improve AI robustness and human-AI team accuracy
• Large-Scale Differentiable Causal Discovery of Factor Graphs
• Near Instance-Optimal PAC Reinforcement Learning for Deterministic MDPs
• A Conditional Randomization Test for Sparse Logistic Regression in High-Dimension
• Learning Distributed and Fair Policies for Network Load Balancing as Markov Potential Game
• On the SDEs and Scaling Rules for Adaptive Gradient Algorithms
• Data Augmentation MCMC for Bayesian Inference from Privatized Data
• Dynamic Tensor Product Regression
• Introspective Learning : A Two-Stage approach for Inference in Neural Networks
• Score-Based Diffusion meets Annealed Importance Sampling
• Local Identifiability of Deep ReLU Neural Networks: the Theory
• Deciding What to Model: Value-Equivalent Sampling for Reinforcement Learning
• A Continuous Time Framework for Discrete Denoising Models
• Are Two Heads the Same as One? Identifying Disparate Treatment in Fair Neural Networks
• Infinite Recommendation Networks: A Data-Centric Approach
• Neural Sheaf Diffusion: A Topological Perspective on Heterophily and Oversmoothing in GNNs
• Learning Distributions Generated by Single-Layer ReLU Networks in the Presence of Arbitrary Outliers
• RainNet: A Large-Scale Imagery Dataset and Benchmark for Spatial Precipitation Downscaling
• VICRegL: Self-Supervised Learning of Local Visual Features
• Learning to Find Proofs and Theorems by Learning to Refine Search Strategies: The Case of Loop Invariant Synthesis
• Generalization for multiclass classification with overparameterized linear models
• Okapi: Generalising Better by Making Statistical Matches Match
• Deterministic Langevin Monte Carlo with Normalizing Flows for Bayesian Inference
• Probabilistic Transformer: Modelling Ambiguities and Distributions for RNA Folding and Molecule Design
• Adversarial Reprogramming Revisited
• A Near-Optimal Best-of-Both-Worlds Algorithm for Online Learning with Feedback Graphs
• Uncertainty Estimation Using Riemannian Model Dynamics for Offline Reinforcement Learning
• Left Heavy Tails and the Effectiveness of the Policy and Value Networks in DNN-based best-first search for Sokoban Planning
• The Pitfalls of Regularization in Off-Policy TD Learning
• OmniVL: One Foundation Model for Image-Language and Video-Language Tasks
• CCCP is Frank-Wolfe in disguise
• Learning-based Motion Planning in Dynamic Environments Using GNNs and Temporal Encoding
• Identifiability and generalizability from multiple experts in Inverse Reinforcement Learning
• Adam Can Converge Without Any Modification On Update Rules
• A Consistent and Differentiable Lp Canonical Calibration Error Estimator
• Exploration-Guided Reward Shaping for Reinforcement Learning under Sparse Rewards
• Detection and Localization of Changes in Conditional Distributions
• TransTab: Learning Transferable Tabular Transformers Across Tables
• Spatial Mixture-of-Experts
• TransBoost: Improving the Best ImageNet Performance using Deep Transduction
• A Multilabel Classification Framework for Approximate Nearest Neighbor Search
• On Efficient Online Imitation Learning via Classification
• Inherently Explainable Reinforcement Learning in Natural Language
• Inverse Game Theory for Stackelberg Games: the Blessing of Bounded Rationality
• Incrementality Bidding via Reinforcement Learning under Mixed and Delayed Rewards
• $k$-Sliced Mutual Information: A Quantitative Study of Scalability with Dimension
• A Direct Approximation of AIXI Using Logical State Abstractions
• Transition to Linearity of General Neural Networks with Directed Acyclic Graph Architecture
• Towards Efficient Post-training Quantization of Pre-trained Language Models
• A Unified Analysis of Federated Learning with Arbitrary Client Participation
• Self-supervised surround-view depth estimation with volumetric feature fusion
• Robust Bayesian Regression via Hard Thresholding
• On the Efficient Implementation of High Accuracy Optimality of Profile Maximum Likelihood
• The price of unfairness in linear bandits with biased feedback
• Cooperative Distribution Alignment via JSD Upper Bound
• Censored Quantile Regression Neural Networks for Distribution-Free Survival Analysis
• Dataset Inference for Self-Supervised Models
• Active Learning Through a Covering Lens
• Adversarially Robust Learning: A Generic Minimax Optimal Learner and Characterization
• Wavelet Score-Based Generative Modeling
• Efficiently Factorizing Boolean Matrices using Proximal Gradient Descent
• The Curse of Unrolling: Rate of Differentiating Through Optimization
• ZeroQuant: Efficient and Affordable Post-Training Quantization for Large-Scale Transformers
• Dual-discriminative Graph Neural Network for Imbalanced Graph-level Anomaly Detection
• Improved Convergence Rate of Stochastic Gradient Langevin Dynamics with Variance Reduction and its Application to Optimization
• What You See is What You Classify: Black Box Attributions
• A Closer Look at Prototype Classifier for Few-shot Image Classification
• Graph Reordering for Cache-Efficient Near Neighbor Search
• Trade-off between Payoff and Model Rewards in Shapley-Fair Collaborative Machine Learning
• Muffliato: Peer-to-Peer Privacy Amplification for Decentralized Optimization and Averaging
• Adaptively Exploiting d-Separators with Causal Bandits
• Generative Neural Articulated Radiance Fields
• Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos
• Identification, Amplification and Measurement: A bridge to Gaussian Differential Privacy
• BagFlip: A Certified Defense Against Data Poisoning
• On the Convergence Theory for Hessian-Free Bilevel Algorithms
• On the Sample Complexity of Stabilizing LTI Systems on a Single Trajectory
• Gradient Descent Is Optimal Under Lower Restricted Secant Inequality And Upper Error Bound
• Maximum Common Subgraph Guided Graph Retrieval: Late and Early Interaction Networks
• Generalized Variational Inference in Function Spaces: Gaussian Measures meet Bayesian Deep Learning
• Refining Low-Resource Unsupervised Translation by Language Disentanglement of Multilingual Translation Model
• Analyzing Data-Centric Properties for Graph Contrastive Learning
• RényiCL: Contrastive Representation Learning with Skew Rényi Divergence
• Scalable Neural Video Representations with Learnable Positional Features
• Towards Improving Calibration in Object Detection Under Domain Shift
• GenSDF: Two-Stage Learning of Generalizable Signed Distance Functions
• Approaching Quartic Convergence Rates for Quasi-Stochastic Approximation with Application to Gradient-Free Optimization
• Neural Circuit Architectural Priors for Embodied Control
• Quality Not Quantity: On the Interaction between Dataset Design and Robustness of CLIP
• Understanding Deep Neural Function Approximation in Reinforcement Learning via $\epsilon$-Greedy Exploration
• LIFT: Language-Interfaced Fine-Tuning for Non-language Machine Learning Tasks
• Does Self-supervised Learning Really Improve Reinforcement Learning from Pixels?
• Stochastic Multiple Target Sampling Gradient Descent
• If Influence Functions are the Answer, Then What is the Question?
• [Re] Replication Study of DECAF: Generating Fair Synthetic Data Using Causally-Aware Generative Networks
• A Projection-free Algorithm for Constrained Stochastic Multi-level Composition Optimization
• A composable machine-learning approach for steady-state simulations on high-resolution grids
• Degradation-Aware Unfolding Half-Shuffle Transformer for Spectral Compressive Imaging
• [Re] Background-Aware Pooling and Noise-Aware Loss for Weakly-Supervised Semantic Segmentation
• Amortized Projection Optimization for Sliced Wasserstein Generative Models
• Trading Off Resource Budgets For Improved Regret Bounds
• Robust On-Policy Sampling for Data-Efficient Policy Evaluation in Reinforcement Learning
• Matching in Multi-arm Bandit with Collision
• Combinatorial Bandits with Linear Constraints: Beyond Knapsacks and Fairness
• Evaluating Graph Generative Models with Contrastively Learned Features
• Single-pass Streaming Lower Bounds for Multi-armed Bandits Exploration with Instance-sensitive Sample Complexity
• The Minority Matters: A Diversity-Promoting Collaborative Metric Learning Algorithm
• A Communication-Efficient Distributed Gradient Clipping Algorithm for Training Deep Neural Networks
• Rate-Distortion Theoretic Bounds on Generalization Error for Distributed Learning
• One-shot Neural Backdoor Erasing via Adversarial Weight Masking
• Learning Generalizable Models for Vehicle Routing Problems via Knowledge Distillation
• Multimodal Contrastive Learning with LIMoE: the Language-Image Mixture of Experts
• Movement Penalized Bayesian Optimization with Application to Wind Energy Systems
• Two-layer neural network on infinite dimensional data: global optimization guarantee in the mean-field regime
• Efficient Aggregated Kernel Tests using Incomplete $U$-statistics
• Recurrent Memory Transformer
• Unsupervised Learning From Incomplete Measurements for Inverse Problems
• An empirical analysis of compute-optimal large language model training
• DIMES: A Differentiable Meta Solver for Combinatorial Optimization Problems
• SHAQ: Incorporating Shapley Value Theory into Multi-Agent Q-Learning
• House of Cans: Covert Transmission of Internal Datasets via Capacity-Aware Neuron Steganography
• A Unifying Framework for Online Optimization with Long-Term Constraints
• Better Best of Both Worlds Bounds for Bandits with Switching Costs
• Renyi Differential Privacy of Propose-Test-Release and Applications to Private and Robust Machine Learning
• Earthformer: Exploring Space-Time Transformers for Earth System Forecasting
• Lower Bounds and Nearly Optimal Algorithms in Distributed Learning with Communication Compression
• Look Around and Refer: 2D Synthetic Semantics Knowledge Distillation for 3D Visual Grounding
• Variational inference via Wasserstein gradient flows
• Efficient Risk-Averse Reinforcement Learning
• Operator Splitting Value Iteration
• Composite Feature Selection Using Deep Ensembles
• From Gradient Flow on Population Loss to Learning with Stochastic Gradient Descent
• Contrastive Adapters for Foundation Model Group Robustness
• Domain Generalization by Learning and Removing Domain-specific Features
• On Uncertainty, Tempering, and Data Augmentation in Bayesian Classification
• Physics-Embedded Neural Networks: Graph Neural PDE Solvers with Mixed Boundary Conditions
• SAVi++: Towards End-to-End Object-Centric Learning from Real-World Videos
• Debiased Self-Training for Semi-Supervised Learning
• Learning Recourse on Instance Environment to Enhance Prediction Accuracy
• Differentially Private Learning with Margin Guarantees
• Provable General Function Class Representation Learning in Multitask Bandits and MDP
• Characterization of Excess Risk for Locally Strongly Convex Population Risk
• Extrapolative Continuous-time Bayesian Neural Network for Fast Training-free Test-time Adaptation
• SNAKE: Shape-aware Neural 3D Keypoint Field
• SIXO: Smoothing Inference with Twisted Objectives
• Learning Articulated Rigid Body Dynamics with Lagrangian Graph Neural Network
• Gradient Descent: The Ultimate Optimizer
• Self-Consistent Dynamical Field Theory of Kernel Evolution in Wide Neural Networks
• Batch size-invariance for policy optimization
• Distributionally robust weighted k-nearest neighbors
• On the Importance of Gradient Norm in PAC-Bayesian Bounds
• Fair Bayes-Optimal Classifiers Under Predictive Parity
• Analyzing Lottery Ticket Hypothesis from PAC-Bayesian Theory Perspective
• Counterfactual Fairness with Partially Known Causal Graph
• When Privacy Meets Partial Information: A Refined Analysis of Differentially Private Bandits
• Efficient identification of informative features in simulation-based inference
• Transform Once: Efficient Operator Learning in Frequency Domain
• Expansion and Shrinkage of Localization for Weakly-Supervised Semantic Segmentation
• Deep Active Learning by Leveraging Training Dynamics
• Rate-Optimal Online Convex Optimization in Adaptive Linear Control
• SAPipe: Staleness-Aware Pipeline for Data Parallel DNN Training
• Understanding Programmatic Weak Supervision via Source-aware Influence Function
• Mind Reader: Reconstructing complex images from brain activities
• A Neural Corpus Indexer for Document Retrieval
• CUP: Critic-Guided Policy Reuse
• Low-Rank Modular Reinforcement Learning via Muscle Synergy
• RORL: Robust Offline Reinforcement Learning via Conservative Smoothing
• Safe Opponent-Exploitation Subgame Refinement
• LAPO: Latent-Variable Advantage-Weighted Policy Optimization for Offline Reinforcement Learning
• A Primer for Neural Arithmetic Logic Modules
• Improving Task-Specific Generalization in Few-Shot Learning via Adaptive Vicinal Risk Minimization
• Chroma-VAE: Mitigating Shortcut Learning with Generative Classifiers
• Look More but Care Less in Video Recognition
• Adversarial Task Up-sampling for Meta-learning
• Let Images Give You More: Point Cloud Cross-Modal Training for Shape Analysis
• Peer Prediction for Learning Agents
• Cache-Augmented Inbatch Importance Resampling for Training Recommender Retriever
• Interaction Modeling with Multiplex Attention
• Learning to Configure Computer Networks with Neural Algorithmic Reasoning
• Can Adversarial Training Be Manipulated By Non-Robust Features?
• Uncertainty-Aware Hierarchical Refinement for Incremental Implicitly-Refined Classification
• MGNNI: Multiscale Graph Neural Networks with Implicit Layers
• Discrete Compositional Representations as an Abstraction for Goal Conditioned Reinforcement Learning
• MoVQ: Modulating Quantized Vectors for High-Fidelity Image Generation
• Learning Mixed Multinomial Logits with Provable Guarantees
• Relational Reasoning via Set Transformers: Provable Efficiency and Applications to MARL
• Hilbert Distillation for Cross-Dimensionality Networks
• Recurrent Video Restoration Transformer with Guided Deformable Attention
• Coarse-to-Fine Vision-Language Pre-training with Fusion in the Backbone
• Unified Optimal Transport Framework for Universal Domain Adaptation
• Learning Deep Input-Output Stable Dynamics
• Batch Bayesian Optimization on Permutations using the Acquisition Weighted Kernel
• Neural Topological Ordering for Computation Graphs
• Memory Efficient Continual Learning with Transformers
• Efficient Knowledge Distillation from Model Checkpoints
• EvenNet: Ignoring Odd-Hop Neighbors Improves Robustness of Graph Neural Networks
• SelecMix: Debiased Learning by Contradicting-pair Sampling
• Coordinate Linear Variance Reduction for Generalized Linear Programming
• Local Latent Space Bayesian Optimization over Structured Inputs
• Memorization and Optimization in Deep Neural Networks with Minimum Over-parameterization
• Debiasing Graph Neural Networks via Learning Disentangled Causal Substructure
• Learning Robust Dynamics through Variational Sparse Gating
• VER: Scaling On-Policy RL Leads to the Emergence of Navigation in Embodied Rearrangement
• A Unified Framework for Deep Symbolic Regression
• [Re] A Cluster-based Approach for Improving Isotropy in Contextual Embedding Space
• Is Sortition Both Representative and Fair?
• All Politics is Local: Redistricting via Local Fairness
• Learning Interface Conditions in Domain Decomposition Solvers
• Off-Policy Evaluation for Action-Dependent Non-stationary Environments
• Factored DRO: Factored Distributionally Robust Policies for Contextual Bandits
• Causal Discovery in Linear Latent Variable Models Subject to Measurement Error
• Human-AI Collaborative Bayesian Optimisation
• SNN-RAT: Robustness-enhanced Spiking Neural Network through Regularized Adversarial Training
• OOD Link Prediction Generalization Capabilities of Message-Passing GNNs in Larger Test Graphs
• Generalized Delayed Feedback Model with Post-Click Information in Recommender Systems
• GAR: Generalized Autoregression for Multi-Fidelity Fusion
• Learning Representations via a Robust Behavioral Metric for Deep Reinforcement Learning
• Environment Diversification with Multi-head Neural Network for Invariant Learning
• MetaTeacher: Coordinating Multi-Model Domain Adaptation for Medical Image Classification
• Collaborative Learning by Detecting Collaboration Partners
• DetCLIP: Dictionary-Enriched Visual-Concept Paralleled Pre-training for Open-world Detection
• Set-based Meta-Interpolation for Few-Task Meta-Learning
• Neural Collapse with Normalized Features: A Geometric Analysis over the Riemannian Manifold
• Error Analysis of Tensor-Train Cross Approximation
• Trading off Utility, Informativeness, and Complexity in Emergent Communication
• Hyper-Representations as Generative Models: Sampling Unseen Neural Network Weights
• Learning to Break the Loop: Analyzing and Mitigating Repetitions for Neural Text Generation
• A Damped Newton Method Achieves Global $\mathcal O \left(\frac{1}{k^2}\right)$ and Local Quadratic Convergence Rate
• Finding Second-Order Stationary Points in Nonconvex-Strongly-Concave Minimax Optimization
• Private Set Generation with Discriminative Information
• Robust Semi-Supervised Learning when Not All Classes have Labels
• Bandit Theory and Thompson Sampling-Guided Directed Evolution for Sequence Optimization
• "Lossless" Compression of Deep Neural Networks: A High-dimensional Neural Tangent Kernel Approach
• GLIF: A Unified Gated Leaky Integrate-and-Fire Neuron for Spiking Neural Networks
• Finding and Listing Front-door Adjustment Sets
• Bridging the Gap from Asymmetry Tricks to Decorrelation Principles in Non-contrastive Self-supervised Learning
• Logical Credal Networks
• Sharp Analysis of Stochastic Optimization under Global Kurdyka-Lojasiewicz Inequality
• SInGE: Sparsity via Integrated Gradients Estimation of Neuron Relevance
• A Robust Phased Elimination Algorithm for Corruption-Tolerant Gaussian Process Bandits
• Rethinking the compositionality of point clouds through regularization in the hyperbolic space
• Identifiability of deep generative models without auxiliary information
• Sub-exponential time Sum-of-Squares lower bounds for Principal Components Analysis
• Robust Anytime Learning of Markov Decision Processes
• COLD Decoding: Energy-based Constrained Text Generation with Langevin Dynamics
• Simultaneous Missing Value Imputation and Structure Learning with Groups
• Provably Efficient Model-Free Constrained RL with Linear Function Approximation
• Private Estimation with Public Data
• Friendly Noise against Adversarial Noise: A Powerful Defense against Data Poisoning Attack
• Multi-Fidelity Best-Arm Identification
• Off-Policy Evaluation with Deficient Support Using Side Information
• Challenging Common Assumptions in Convex Reinforcement Learning
• Decision Trees with Short Explainable Rules
• List-Decodable Sparse Mean Estimation
• Stochastic Adaptive Activation Function
• Rethinking Knowledge Graph Evaluation Under the Open-World Assumption
• A Theoretical Framework for Inference Learning
• OPEN: Orthogonal Propagation with Ego-Network Modeling
• On the Frequency-bias of Coordinate-MLPs
• Generalization Properties of NAS under Activation and Skip Connection Search
• Robustness in deep learning: The good (width), the bad (depth), and the ugly (initialization)
• Extrapolation and Spectral Bias of Neural Nets with Hadamard Product: a Polynomial Net Study
• A Rotated Hyperbolic Wrapped Normal Distribution for Hierarchical Representation Learning
• Controllable 3D Face Synthesis with Conditional Generative Occupancy Fields
• A general approximation lower bound in $L^p$ norm, with applications to feed-forward neural networks
• CLOOB: Modern Hopfield Networks with InfoLOOB Outperform CLIP
• Communication Efficient Distributed Learning for Kernelized Contextual Bandits
• Communication Efficient Federated Learning for Generalized Linear Bandits
• Versatile Multi-stage Graph Neural Network for Circuit Representation
• Cost-Sensitive Self-Training for Optimizing Non-Decomposable Metrics
• Understanding Square Loss in Training Overparametrized Neural Network Classifiers
• The Gyro-Structure of Some Matrix Manifolds
• Multi-view Subspace Clustering on Topological Manifold
• HyperDomainNet: Universal Domain Adaptation for Generative Adversarial Networks
• S$^3$-NeRF: Neural Reflectance Field from Shading and Shadow under a Single Viewpoint
• PaCo: Parameter-Compositional Multi-task Reinforcement Learning
• IALE: Imitating Active Learner Ensembles
• Score-based Generative Modeling Secretly Minimizes the Wasserstein Distance
• Estimating Noise Transition Matrix with Label Correlations for Noisy Multi-Label Learning
• Momentum Aggregation for Private Non-convex ERM
• SCL-WC: Cross-Slide Contrastive Learning for Weakly-Supervised Whole-Slide Image Classification
• Differentially Private Online-to-batch for Smooth Losses
• Distributionally Robust Optimization with Data Geometry
• Decentralized Training of Foundation Models in Heterogeneous Environments
• On the convergence of policy gradient methods to Nash equilibria in general stochastic games
• Sample Complexity of Learning Heuristic Functions for Greedy-Best-First and A* Search
• Rank Diminishing in Deep Neural Networks
• Don't Pour Cereal into Coffee: Differentiable Temporal Logic for Temporal Action Segmentation
• Lethal Dose Conjecture on Data Poisoning
• Learning Substructure Invariance for Out-of-Distribution Molecular Representations
• NeuPhysics: Editable Neural Geometry and Physics from Monocular Videos
• Understanding the Evolution of Linear Regions in Deep Reinforcement Learning
• RecursiveMix: Mixed Learning with History
• DeepTOP: Deep Threshold-Optimal Policy for MDPs and RMABs
• Fairness Reprogramming
• S-Prompts Learning with Pre-trained Transformers: An Occam’s Razor for Domain Incremental Learning
• Coded Residual Transform for Generalizable Deep Metric Learning
• Embodied Scene-aware Human Pose Estimation
• Generative Status Estimation and Information Decoupling for Image Rain Removal
• Subsidiary Prototype Alignment for Universal Domain Adaptation
• Align then Fusion: Generalized Large-scale Multi-view Clustering with Anchor Matching Correspondences
• DOMINO: Decomposed Mutual Information Optimization for Generalized Context in Meta-Reinforcement Learning
• EcoFormer: Energy-Saving Attention with Linear Complexity
• Machine Learning on Graphs: A Model and Comprehensive Taxonomy
• DAGMA: Learning DAGs via M-matrices and a Log-Determinant Acyclicity Characterization
• Self-Supervised Visual Representation Learning with Semantic Grouping
• Mind the Gap: Understanding the Modality Gap in Multi-modal Contrastive Representation Learning
• Practical Adversarial Attacks on Spatiotemporal Traffic Forecasting Models
• Active Labeling: Streaming Stochastic Gradients
• SegNeXt: Rethinking Convolutional Attention Design for Semantic Segmentation
• Polynomial Neural Fields for Subband Decomposition and Manipulation
• Visual Concepts Tokenization
• Phase Transition from Clean Training to Adversarial Training
• HSurf-Net: Normal Estimation for 3D Point Clouds by Learning Hyper Surfaces
• Adversarial Style Augmentation for Domain Generalized Urban-Scene Segmentation
• Natural Color Fool: Towards Boosting Black-box Unrestricted Attacks
• Cross-Image Context for Single Image Inpainting
• TOIST: Task Oriented Instance Segmentation Transformer with Noun-Pronoun Distillation
• Is Out-of-Distribution Detection Learnable?
• Masked Autoencoders As Spatiotemporal Learners
• Trap and Replace: Defending Backdoor Attacks by Trapping Them into an Easy-to-Replace Subnetwork
• PolarMix: A General Data Augmentation Technique for LiDAR Point Clouds
• Prototypical VoteNet for Few-Shot 3D Point Cloud Object Detection
• Optimistic Tree Searches for Combinatorial Black-Box Optimization
• Tensor Wheel Decomposition and Its Tensor Completion Application
• PALBERT: Teaching ALBERT to Ponder
• Towards Efficient 3D Object Detection with Knowledge Distillation
• Towards Lightweight Black-Box Attack Against Deep Neural Networks
• HumanLiker: A Human-like Object Detector to Model the Manual Labeling Process
• Learn what matters: cross-domain imitation learning with task-relevant embeddings
• Whitening Convergence Rate of Coupling-based Normalizing Flows
• Hierarchical Normalization for Robust Monocular Depth Estimation
• Unsupervised Multi-Object Segmentation by Predicting Probable Motion Patterns
• On the Strong Correlation Between Model Invariance and Generalization
• Q-ViT: Accurate and Fully Quantized Low-bit Vision Transformer
• Fully Sparse 3D Object Detection
• Learning Multi-resolution Functional Maps with Spectral Attention for Robust Shape Matching
• A Coupled Design of Exploiting Record Similarity for Practical Vertical Federated Learning
• Towards Robust Blind Face Restoration with Codebook Lookup Transformer
• Improved Fine-Tuning by Better Leveraging Pre-Training Data
• TotalSelfScan: Learning Full-body Avatars from Self-Portrait Videos of Faces, Hands, and Bodies
• Cross Aggregation Transformer for Image Restoration
• Behavior Transformers: Cloning $k$ modes with one stone
• What Makes a "Good" Data Augmentation in Knowledge Distillation - A Statistical Perspective
• Bridging the Gap between Object and Image-level Representations for Open-Vocabulary Detection
• Discrete-Convex-Analysis-Based Framework for Warm-Starting Algorithms with Predictions
• Divert More Attention to Vision-Language Tracking
• Trajectory Inference via Mean-field Langevin in Path Space
• ElasticMVS: Learning elastic part representation for self-supervised multi-view stereopsis
• A2: Efficient Automated Attacker for Boosting Adversarial Training
• PerfectDou: Dominating DouDizhu with Perfect Information Distillation
• MsSVT: Mixed-scale Sparse Voxel Transformer for 3D Object Detection on Point Clouds
• Towards Versatile Embodied Navigation
• Product Ranking for Revenue Maximization with Multiple Purchases
• Remember the Past: Distilling Datasets into Addressable Memories for Neural Networks
• ResT V2: Simpler, Faster and Stronger
• In the Eye of the Beholder: Robust Prediction with Causal User Modeling
• Bi-directional Weakly Supervised Knowledge Distillation for Whole Slide Image Classification
• Multi-modal Grouping Network for Weakly-Supervised Audio-Visual Video Parsing
• Mining Unseen Classes via Regional Objectness: A Simple Baseline for Incremental Segmentation
• Panchromatic and Multispectral Image Fusion via Alternating Reverse Filtering Network
• Pay attention to your loss : understanding misconceptions about Lipschitz neural networks
• End-to-end Symbolic Regression with Transformers
• SPoVT: Semantic-Prototype Variational Transformer for Dense Point Cloud Semantic Completion
• Unsupervised Representation Learning from Pre-trained Diffusion Probabilistic Models
• What I Cannot Predict, I Do Not Understand: A Human-Centered Evaluation Framework for Explainability Methods
• Stochastic Window Transformer for Image Restoration
• A Closer Look at Weakly-Supervised Audio-Visual Source Localization
• Semi-Discrete Normalizing Flows through Differentiable Tessellation
• Blackbox Attacks via Surrogate Ensemble Search
• Saliency-Aware Neural Architecture Search
• ViTPose: Simple Vision Transformer Baselines for Human Pose Estimation
• Learning Best Combination for Efficient N:M Sparsity
• Predicting Label Distribution from Multi-label Ranking
• Instance-Based Uncertainty Estimation for Gradient-Boosted Regression Trees
• Semantic Diffusion Network for Semantic Segmentation
• Regret Bounds for Information-Directed Reinforcement Learning
• A Spectral Approach to Item Response Theory
• UDC: Unified DNAS for Compressible TinyML Models for Neural Processing Units
• AutoLink: Self-supervised Learning of Human Skeletons and Object Outlines by Linking Keypoints
• Optimistic Mirror Descent Either Converges to Nash or to Strong Coarse Correlated Equilibria in Bimatrix Games
• Parameter-Efficient Masking Networks
• Learning Distinct and Representative Modes for Image Captioning
• Fully Convolutional One-Stage 3D Object Detection on LiDAR Range Images
• HUMANISE: Language-conditioned Human Motion Generation in 3D Scenes
• VCT: A Video Compression Transformer
• Non-stationary Transformers: Exploring the Stationarity in Time Series Forecasting
• VITA: Video Instance Segmentation via Object Token Association
• A Unified Analysis of Mixed Sample Data Augmentation: A Loss Function Perspective
• Geometry-aware Two-scale PIFu Representation for Human Reconstruction
• Causally motivated multi-shortcut identification and removal
• SegViT: Semantic Segmentation with Plain Vision Transformers
• Inducing Neural Collapse in Imbalanced Learning: Do We Really Need a Learnable Classifier at the End of Deep Neural Network?
• Masked Autoencoders that Listen
• Semi-Supervised Semantic Segmentation via Gentle Teaching Assistant
• Video-based Human-Object Interaction Detection from Tubelet Tokens
• Learning Equivariant Segmentation with Instance-Unique Querying
• Enhanced Latent Space Blind Model for Real Image Denoising via Alternative Optimization
• High-dimensional Additive Gaussian Processes under Monotonicity Constraints
• Learning Generalizable Part-based Feature Representation for 3D Point Clouds
• Constants of motion network
• Asymptotically Unbiased Instance-wise Regularized Partial AUC Optimization: Theory and Algorithm
• Rethinking Alignment in Video Super-Resolution Transformers
• Robust Testing in High-Dimensional Sparse Models
• INRAS: Implicit Neural Representation for Audio Scenes
• BMU-MoCo: Bidirectional Momentum Update for Continual Video-Language Modeling
• DropCov: A Simple yet Effective Method for Improving Deep Architectures
• Antigen-Specific Antibody Design and Optimization with Diffusion-Based Generative Models for Protein Structures
• Monocular Dynamic View Synthesis: A Reality Check
• A Mixture Of Surprises for Unsupervised Reinforcement Learning
• QueryPose: Sparse Multi-Person Pose Regression via Spatial-Aware Part-Level Query
• Decoupling Knowledge from Memorization: Retrieval-augmented Prompt Learning
• Misspecified Phase Retrieval with Generative Priors
• Watermarking for Out-of-distribution Detection
• Error Correction Code Transformer
• Maximum Class Separation as Inductive Bias in One Matrix
• Sequencer: Deep LSTM for Image Classification
• Self-Supervised Learning via Maximum Entropy Coding
• Giga-scale Kernel Matrix-Vector Multiplication on GPU
• Scalable Infomin Learning
• Multi-dataset Training of Transformers for Robust Action Recognition
• ZARTS: On Zero-order Optimization for Neural Architecture Search
• Online Training Through Time for Spiking Neural Networks
• Multi-Instance Causal Representation Learning for Instance Label Prediction and Out-of-Distribution Generalization
• P2P: Tuning Pre-trained Image Models for Point Cloud Analysis with Point-to-Pixel Prompting
• Towards Theoretically Inspired Neural Initialization Optimization
• Vision GNN: An Image is Worth Graph of Nodes
• Rotation-Equivariant Conditional Spherical Neural Fields for Learning a Natural Illumination Prior
• Supported Policy Optimization for Offline Reinforcement Learning
• AutoMS: Automatic Model Selection for Novelty Detection with Error Rate Control
• Increasing Confidence in Adversarial Robustness Evaluations
• Generalization Bounds for Estimating Causal Effects of Continuous Treatments
• Non-Markovian Reward Modelling from Trajectory Labels via Interpretable Multiple Instance Learning
• Learning Consistency-Aware Unsigned Distance Functions Progressively from Raw Point Clouds
• Why Do Artificially Generated Data Help Adversarial Robustness
• Learning Infinite-Horizon Average-Reward Restless Multi-Action Bandits via Index Awareness
• Theory and Approximate Solvers for Branched Optimal Transport with Multiple Sources
• New Lower Bounds for Private Estimation and a Generalized Fingerprinting Lemma
• PointTAD: Multi-Label Temporal Action Detection with Learnable Query Points
• On the Generalizability and Predictability of Recommender Systems
• Polyhistor: Parameter-Efficient Multi-Task Adaptation for Dense Vision Tasks
• Generative Visual Prompt: Unifying Distributional Control of Pre-Trained Generative Models
• Stability and Generalization Analysis of Gradient Methods for Shallow Neural Networks
• Physically-Based Face Rendering for NIR-VIS Face Recognition
• Adversarial Training with Complementary Labels: On the Benefit of Gradually Informative Attacks
• Few-Shot Continual Active Learning by a Robot
• MultiScan: Scalable RGBD scanning for 3D environments with articulated objects
• Transformer-based Working Memory for Multiagent Reinforcement Learning with Action Parsing
• Structural Kernel Search via Bayesian Optimization and Symbolical Optimal Transport
• Biologically Inspired Dynamic Thresholds for Spiking Neural Networks
• Don't Roll the Dice, Ask Twice: The Two-Query Distortion of Matching Problems and Beyond
• A Unified Model for Multi-class Anomaly Detection
• A framework for bilevel optimization that enables stochastic and global variance reduction algorithms
• SAViT: Structure-Aware Vision Transformer Pruning via Collaborative Optimization
• Masked Generative Adversarial Networks are Data-Efficient Generation Learners
• Training Spiking Neural Networks with Event-driven Backpropagation
• MCMAE: Masked Convolution Meets Masked Autoencoders
• Learning Physical Dynamics with Subequivariant Graph Neural Networks
• Online PAC-Bayes Learning
• Implicit Warping for Animation with Image Sets
• Rethinking Resolution in the Context of Efficient Video Recognition
• RAMBO-RL: Robust Adversarial Model-Based Offline Reinforcement Learning
• CEBaB: Estimating the Causal Effects of Real-World Concepts on NLP Model Behavior
• Natural gradient enables fast sampling in spiking neural networks
• MultiGuard: Provably Robust Multi-label Classification against Adversarial Examples
• Efficient and Effective Multi-task Grouping via Meta Learning on Task Combinations
• Robust Calibration with Multi-domain Temperature Scaling
• Exploration via Planning for Information about the Optimal Trajectory
• Mean Estimation in High-Dimensional Binary Markov Gaussian Mixture Models
• BiT: Robustly Binarized Multi-distilled Transformer
• PopArt: Efficient Sparse Regression and Experimental Design for Optimal Sparse Linear Bandits
• On the Effect of Pre-training for Transformer in Different Modality on Offline Reinforcement Learning
• On-Device Training Under 256KB Memory
• Geo-SIC: Learning Deformable Geometric Shapes in Deep Image Classifiers
• An Embarrassingly Simple Approach to Semi-Supervised Few-Shot Learning
• Multi-Granularity Cross-modal Alignment for Generalized Medical Visual Representation Learning
• Finite-Time Analysis of Adaptive Temporal Difference Learning with Deep Neural Networks
• Neural Surface Reconstruction of Dynamic Scenes with Monocular RGB-D Camera
• Mutual Information Divergence: A Unified Metric for Multimodal Generative Models
• Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding
• GLIPv2: Unifying Localization and Vision-Language Understanding
• A Unified Diversity Measure for Multiagent Reinforcement Learning
• Plan To Predict: Learning an Uncertainty-Foreseeing Model For Model-Based Reinforcement Learning
• Learning to Accelerate Partial Differential Equations via Latent Global Evolution
• Active Learning for Multiple Target Models
• Alignment-guided Temporal Attention for Video Action Recognition
• Open-Ended Reinforcement Learning with Neural Reward Functions
• On Margins and Generalisation for Voting Classifiers
• Contrastive Neural Ratio Estimation
• Mildly Conservative Q-Learning for Offline Reinforcement Learning
• Self-Supervised Image Restoration with Blurry and Noisy Pairs
• Recommender Forest for Efficient Retrieval
• Retrieval-Augmented Diffusion Models
• PatchComplete: Learning Multi-Resolution Patch Priors for 3D Shape Completion on Unseen Categories
• Generalized Laplacian Eigenmaps
• SAPA: Similarity-Aware Point Affiliation for Feature Upsampling
• Expediting Large-Scale Vision Transformer for Dense Prediction without Fine-tuning
• Random Sharpness-Aware Minimization
• Generalized One-shot Domain Adaptation of Generative Adversarial Networks
• SCONE: Surface Coverage Optimization in Unknown Environments by Volumetric Integration
• A Quantitative Geometric Approach to Neural-Network Smoothness
• Is this the Right Neighborhood? Accurate and Query Efficient Model Agnostic Explanations
• Parametrically Retargetable Decision-Makers Tend To Seek Power
• Learning Individualized Treatment Rules with Many Treatments: A Supervised Clustering Approach Using Adaptive Fusion
• Differentially Private Model Compression
• Is a Modular Architecture Enough?
• Learning General World Models in a Handful of Reward-Free Deployments
• Revisiting Heterophily For Graph Neural Networks
• Recipe for a General, Powerful, Scalable Graph Transformer
• CEIP: Combining Explicit and Implicit Priors for Reinforcement Learning with Demonstrations
• GhostNetV2: Enhance Cheap Operation with Long-Range Attention
• Elucidating the Design Space of Diffusion-Based Generative Models
• Intermediate Prototype Mining Transformer for Few-Shot Semantic Segmentation
• Robust Models are less Over-Confident
• OST: Improving Generalization of DeepFake Detection via One-Shot Test-Time Training
• KSD Aggregated Goodness-of-fit Test
• Distributional Reward Estimation for Effective Multi-agent Deep Reinforcement Learning
• A Near-Optimal Primal-Dual Method for Off-Policy Learning in CMDP
• ZIN: When and How to Learn Invariance Without Environment Partition?
• Enhance the Visual Representation via Discrete Adversarial Training
• Frank-Wolfe-based Algorithms for Approximating Tyler's M-estimator
• Spending Thinking Time Wisely: Accelerating MCTS with Virtual Expansions
• Efficient and Effective Optimal Transport-Based Biclustering
• SageMix: Saliency-Guided Mixup for Point Clouds
• Heatmap Distribution Matching for Human Pose Estimation
• Autoregressive Search Engines: Generating Substrings as Document Identifiers
• Mirror Descent with Relative Smoothness in Measure Spaces, with application to Sinkhorn and EM
• Deconfounded Representation Similarity for Comparison of Neural Networks
• Rethinking Lipschitz Neural Networks and Certified Robustness: A Boolean Function Perspective
• Fine-Grained Analysis of Stability and Generalization for Modern Meta Learning Algorithms
• Gold-standard solutions to the Schrödinger equation using deep learning: How much physics do we need?
• Adv-Attribute: Inconspicuous and Transferable Adversarial Attack on Face Recognition
• Out-of-Distribution Detection with An Adaptive Likelihood Ratio on Informative Hierarchical VAE
• Relational Proxies: Emergent Relationships as Fine-Grained Discriminators
• Sampling without Replacement Leads to Faster Rates in Finite-Sum Minimax Optimization
• Unsupervised Cross-Task Generalization via Retrieval Augmentation
• coVariance Neural Networks
• On the inability of Gaussian process regression to optimally learn compositional functions
• Distributed Methods with Compressed Communication for Solving Variational Inequalities, with Theoretical Guarantees
• When to Update Your Model: Constrained Model-based Reinforcement Learning
• Bayesian Optimization over Discrete and Mixed Spaces via Probabilistic Reparameterization
• Constrained Langevin Algorithms with L-mixing External Random Variables
• Practical Adversarial Multivalid Conformal Prediction
• Biologically-Plausible Determinant Maximization Neural Networks for Blind Separation of Correlated Sources
• Deep Generalized Schrödinger Bridge
• Deep Generative Model for Periodic Graphs
• Optimal Comparator Adaptive Online Learning with Switching Cost
• Enhanced Bilevel Optimization via Bregman Distance
• Learning State-Aware Visual Representations from Audible Interactions
• Near-Optimal Multi-Agent Learning for Safe Coverage Control
• Probabilistic Missing Value Imputation for Mixed Categorical and Ordered Data
• Exploration via Elliptical Episodic Bonuses
• GAUDI: A Neural Architect for Immersive 3D Scene Generation
• Periodic Graph Transformers for Crystal Material Property Prediction
• Parallel Tempering With a Variational Reference
• On the consistent estimation of optimal Receiver Operating Characteristic (ROC) curve
• NS3: Neuro-symbolic Semantic Code Search
• A Deep Learning Dataloader with Shared Data Preparation
• Deep Multi-Modal Structural Equations For Causal Effect Estimation With Unstructured Proxies
• Improving Variational Autoencoders with Density Gap-based Regularization
• Fused Orthogonal Alternating Least Squares for Tensor Clustering
• Representing Spatial Trajectories as Distributions
• Model-Based Offline Reinforcement Learning with Pessimism-Modulated Dynamics Belief
• CLEAR: Generative Counterfactual Explanations on Graphs
• Wasserstein $K$-means for clustering probability distributions
• Biologically-plausible backpropagation through arbitrary timespans via local neuromodulators
• Cost-efficient Gaussian tensor network embeddings for tensor-structured inputs
• Hub-Pathway: Transfer Learning from A Hub of Pre-trained Models
• Green Hierarchical Vision Transformer for Masked Image Modeling
• Beyond the Best: Distribution Functional Estimation in Infinite-Armed Bandits
• An Investigation into Whitening Loss for Self-supervised Learning
• Fixed-Distance Hamiltonian Monte Carlo
• SecureFedYJ: a safe feature Gaussianization protocol for Federated Learning
• Category-Level 6D Object Pose Estimation in the Wild: A Semi-Supervised Learning Approach and A New Dataset
• Deep Attentive Belief Propagation: Integrating Reasoning and Learning for Solving Constraint Optimization Problems
• Amortized Mixing Coupling Processes for Clustering
• HorNet: Efficient High-Order Spatial Interactions with Recursive Gated Convolutions
• Weakly supervised causal representation learning
• Less-forgetting Multi-lingual Fine-tuning
• Online Convex Optimization with Hard Constraints: Towards the Best of Two Worlds and Beyond
• Rethinking Variational Inference for Probabilistic Programs with Stochastic Support
• Retrieve, Reason, and Refine: Generating Accurate and Faithful Patient Instructions
• Cross-modal Learning for Image-Guided Point Cloud Shape Completion
• TCT: Convexifying Federated Learning using Bootstrapped Neural Tangent Kernels
• FNeVR: Neural Volume Rendering for Face Animation
• Bessel Equivariant Networks for Inversion of Transmission Effects in Multi-Mode Optical Fibres
• Bidirectional Learning for Offline Infinite-width Model-based Optimization
• TREC: Transient Redundancy Elimination-based Convolution
• DivBO: Diversity-aware CASH for Ensemble Learning
• Forecasting Human Trajectory from Scene History
• Wasserstein Logistic Regression with Mixed Features
• Contextual Bandits with Knapsacks for a Conversion Model
• Diagnosing failures of fairness transfer across distribution shift in real-world medical settings
• Adaptation Accelerating Sampling-based Bayesian Inference in Attractor Neural Networks
• ELASTIC: Numerical Reasoning with Adaptive Symbolic Compiler
• Oscillatory Tracking of Continuous Attractor Neural Networks Account for Phase Precession and Procession of Hippocampal Place Cells
• UQGAN: A Unified Model for Uncertainty Quantification of Deep Classifiers trained via Conditional GANs
• Make Sharpness-Aware Minimization Stronger: A Sparsified Perturbation Approach
• Contrastive Learning as Goal-Conditioned Reinforcement Learning
• Learning Viewpoint-Agnostic Visual Representations by Recovering Tokens in 3D Space
• When are Local Queries Useful for Robust Learning?
• Shield Decentralization for Safe Multi-Agent Reinforcement Learning
• Extracting computational mechanisms from neural data using low-rank RNNs
• Data Distributional Properties Drive Emergent In-Context Learning in Transformers
• A Quadrature Rule combining Control Variates and Adaptive Importance Sampling
• Dynamic Fair Division with Partial Information
• Improved Imaging by Invex Regularizers with Global Optima Guarantees
• Markov Chain Score Ascent: A Unifying Framework of Variational Inference with Markovian Gradients
• Change-point Detection for Sparse and Dense Functional Data in General Dimensions
• Sample-Efficient Learning of Correlated Equilibria in Extensive-Form Games
• Local Spatiotemporal Representation Learning for Longitudinally-consistent Neuroimage Analysis
• Parameter tuning and model selection in Optimal Transport with semi-dual Brenier formulation
• Policy Optimization with Advantage Regularization for Long-Term Fairness in Decision Systems
• Generating multivariate time series with COmmon Source CoordInated GAN (COSCI-GAN)
• Unsupervised Multi-View Object Segmentation Using Radiance Field Propagation
• Online Deep Equilibrium Learning for Regularization by Denoising
• When does return-conditioned supervised learning work for offline reinforcement learning?
• Inductive Logical Query Answering in Knowledge Graphs
• The Unreliability of Explanations in Few-shot Prompting for Textual Reasoning
• Biological Learning of Irreducible Representations of Commuting Transformations
• The price of ignorance: how much does it cost to forget noise structure in low-rank matrix estimation?
• MCVD - Masked Conditional Video Diffusion for Prediction, Generation, and Interpolation
• Bayesian Clustering of Neural Spiking Activity Using a Mixture of Dynamic Poisson Factor Analyzers
• Non-convex online learning via algorithmic equivalence
• Decomposing NeRF for Editing via Feature Field Distillation
• Approximate Value Equivalence
• Neur2SP: Neural Two-Stage Stochastic Programming
• Memorization Without Overfitting: Analyzing the Training Dynamics of Large Language Models
• SparCL: Sparse Continual Learning on the Edge
• Visual Prompting via Image Inpainting
• Test-Time Training with Masked Autoencoders
• Visual Clues: Bridging Vision and Language Foundations for Image Paragraph Captioning
• Gradient-Free Methods for Deterministic and Stochastic Nonsmooth Nonconvex Optimization
• BILCO: An Efficient Algorithm for Joint Alignment of Time Series
• Hardness of Noise-Free Learning for Two-Hidden-Layer Neural Networks
• Falsification before Extrapolation in Causal Effect Estimation
• LION: Latent Point Diffusion Models for 3D Shape Generation
• FedRolex: Model-Heterogeneous Federated Learning with Rolling Sub-Model Extraction
• Implicit Regularization or Implicit Conditioning? Exact Risk Trajectories of SGD in High Dimensions
• Sharpness-Aware Training for Free
• CHIMLE: Conditional Hierarchical IMLE for Multimodal Conditional Image Synthesis
• 3DILG: Irregular Latent Grids for 3D Generative Modeling
• Translation-equivariant Representation in Recurrent Networks with a Continuous Manifold of Attractors
• Optimal Transport-based Identity Matching for Identity-invariant Facial Expression Recognition
• Towards Learning Universal Hyperparameter Optimizers with Transformers
• OrdinalCLIP: Learning Rank Prompts for Language-Guided Ordinal Regression
• ComGAN: Unsupervised Disentanglement and Segmentation via Image Composition
• Non-Linear Coordination Graphs
• Towards Hard-pose Virtual Try-on via 3D-aware Global Correspondence Learning
• Fast Distance Oracles for Any Symmetric Norm
• Low-rank Optimal Transport: Approximation, Statistics and Debiasing
• Iterative Scene Graph Generation
• Eliciting Thinking Hierarchy without a Prior
• Learning Robust Rule Representations for Abstract Reasoning via Internal Inferences
• Multi-layer State Evolution Under Random Convolutional Design
• Latency-aware Spatial-wise Dynamic Networks
• Margin-Based Few-Shot Class-Incremental Learning with Class-Level Overfitting Mitigation
• Relation-Constrained Decoding for Text Generation
• Searching for Better Spatio-temporal Alignment in Few-Shot Action Recognition
• Could Giant Pre-trained Image Models Extract Universal Representations?
• IM-Loss: Information Maximization Loss for Spiking Neural Networks
• TokenMixup: Efficient Attention-guided Token-level Data Augmentation for Transformers
• Hyperbolic Feature Augmentation via Distribution Estimation and Infinite Sampling on Manifolds
• Verification and search algorithms for causal DAGs
• AD-DROP: Attribution-Driven Dropout for Robust Language Model Fine-Tuning
• Where to Pay Attention in Sparse Training for Feature Selection?
• TA-MoE: Topology-Aware Large Scale Mixture-of-Expert Training
• Understanding the Failure of Batch Normalization for Transformers in NLP
• Transformers meet Stochastic Block Models: Attention with Data-Adaptive Sparsity and Cost
• Theoretically Provable Spiking Neural Networks
• Deep Combinatorial Aggregation
• Transcormer: Transformer for Sentence Scoring with Sliding Language Modeling
• Self-Supervised Learning with an Information Maximization Criterion
• Improved Utility Analysis of Private CountSketch
• A Classification of $G$-invariant Shallow Neural Networks
• Module-Aware Optimization for Auxiliary Learning
• Incorporating Bias-aware Margins into Contrastive Loss for Collaborative Filtering
• Riemannian Score-Based Generative Modelling
• Out-of-Distribution Detection via Conditional Kernel Independence Model
• Towards Effective Multi-Modal Interchanges in Zero-Resource Sounding Object Localization
• Policy Gradient With Serial Markov Chain Reasoning
• Estimating graphical models for count data with applications to single-cell gene network
• Improving 3D-aware Image Synthesis with A Geometry-aware Discriminator
• Egocentric Video-Language Pretraining
• Efficient Submodular Optimization under Noise: Local Search is Robust
• Conservative Dual Policy Optimization for Efficient Model-Based Reinforcement Learning
• DualCoOp: Fast Adaptation to Multi-Label Recognition with Limited Annotations
• Does Momentum Change the Implicit Regularization on Separable Data?
• VRL3: A Data-Driven Framework for Visual Deep Reinforcement Learning
• Exact Solutions of a Deep Linear Network
• ST-Adapter: Parameter-Efficient Image-to-Video Transfer Learning
• Masked Prediction: A Parameter Identifiability View
• Direct Advantage Estimation
• Depth is More Powerful than Width with Prediction Concatenation in Deep Forest
• AniFaceGAN: Animatable 3D-Aware Face Image Generation for Video Avatars
• Instance-based Learning for Knowledge Base Completion
• Efficient and Effective Augmentation Strategy for Adversarial Training
• u-HuBERT: Unified Mixed-Modal Speech Pretraining And Zero-Shot Transfer to Unlabeled Modality
• First-Order Algorithms for Min-Max Optimization in Geodesic Metric Spaces
• GENIE: Higher-Order Denoising Diffusion Solvers
• Scalable and Efficient Training of Large Convolutional Neural Networks with Differential Privacy
• Structured Recognition for Generative Models with Explaining Away
• UniCLIP: Unified Framework for Contrastive Language-Image Pre-training
• InsNet: An Efficient, Flexible, and Performant Insertion-based Text Generation Model
• Local-Global MCMC kernels: the best of both worlds
• Manifold Interpolating Optimal-Transport Flows for Trajectory Inference
• Doubly Robust Counterfactual Classification
• Uncertainty-Aware Reinforcement Learning for Risk-Sensitive Player Evaluation in Sports Game
• Smooth Fictitious Play in Stochastic Games with Perturbed Payoffs and Unknown Transitions
• SKFlow: Learning Optical Flow with Super Kernels
• Non-stationary Bandits with Knapsacks
• Weighted Mutual Learning with Diversity-Driven Model Compression
• Learning to Attack Federated Learning: A Model-based Reinforcement Learning Attack Framework
• Improving Zero-Shot Generalization in Offline Reinforcement Learning using Generalized Similarity Functions
• Procedural Image Programs for Representation Learning
• Bivariate Causal Discovery for Categorical Data via Classification with Optimal Label Permutation
• High-Order Pooling for Graph Neural Networks with Tensor Decomposition
• TTOpt: A Maximum Volume Quantized Tensor Train-based Optimization and its Application to Reinforcement Learning
• SALSA: Attacking Lattice Cryptography with Transformers
• Class-Aware Adversarial Transformers for Medical Image Segmentation
• A Single-timescale Analysis for Stochastic Approximation with Multiple Coupled Sequences
• You Only Live Once: Single-Life Reinforcement Learning
• Semi-Supervised Learning with Decision Trees: Graph Laplacian Tree Alternating Optimization
• When does dough become a bagel? Analyzing the remaining mistakes on ImageNet
• Learning from Stochastically Revealed Preference
• A Best-of-Both-Worlds Algorithm for Bandits with Delayed Feedback
• Online Minimax Multiobjective Optimization: Multicalibeating and Other Applications
• Algorithms that Approximate Data Removal: New Results and Limitations
• Annihilation of Spurious Minima in Two-Layer ReLU Networks
• Unsupervised Image-to-Image Translation with Density Changing Regularization
• Reproducibility in Optimization: Theoretical Framework and Limits
• Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering
• Systematic improvement of neural network quantum states using Lanczos
• Privacy of Noisy Stochastic Gradient Descent: More Iterations without More Privacy Loss
• Diagonal State Spaces are as Effective as Structured State Spaces
• Why neural networks find simple solutions: The many regularizers of geometric complexity
• Zero-Shot 3D Drug Design by Sketching and Generating
• Adaptive Oracle-Efficient Online Learning
• Brownian Noise Reduction: Maximizing Privacy Subject to Accuracy Constraints
• Efficient Active Learning with Abstention
• Unsupervised Learning of Shape Programs with Repeatable Implicit Parts
• Moderate-fitting as a Natural Backdoor Defender for Pre-trained Language Models
• Controllable Text Generation with Neurally-Decomposed Oracle
• A Fast Post-Training Pruning Framework for Transformers
• ConfounderGAN: Protecting Image Data Privacy with Causal Confounder
• Improved Feature Distillation via Projector Ensemble
• Neuron with Steady Response Leads to Better Generalization
• Mirror Descent Maximizes Generalized Margin and Can Be Implemented Efficiently
• Self-Organized Group for Cooperative Multi-agent Reinforcement Learning
• APG: Adaptive Parameter Generation Network for Click-Through Rate Prediction
• Learning Manifold Dimensions with Conditional Variational Autoencoders
• Discovering Design Concepts for CAD Sketches
• Reconstruction on Trees and Low-Degree Polynomials
• Test Time Adaptation via Conjugate Pseudo-labels
• Pre-Trained Image Encoder for Generalizable Visual Reinforcement Learning
• GenerSpeech: Towards Style Transfer for Generalizable Out-Of-Domain Text-to-Speech
• Momentum Adversarial Distillation: Handling Large Distribution Shifts in Data-Free Knowledge Distillation
• FreGAN: Exploiting Frequency Components for Training GANs under Limited Data
• FasterRisk: Fast and Accurate Interpretable Risk Scores
• When to Trust Your Simulator: Dynamics-Aware Hybrid Offline-and-Online Reinforcement Learning
• Generalization Bounds for Stochastic Gradient Descent via Localized $\varepsilon$-Covers
• Symbolic Distillation for Learned TCP Congestion Control
• Proximal Learning With Opponent-Learning Awareness
• Accelerated Linearized Laplace Approximation for Bayesian Deep Learning
• GAGA: Deciphering Age-path of Generalized Self-paced Regularizer
• Provable Benefit of Multitask Representation Learning in Reinforcement Learning
• Follow-the-Perturbed-Leader for Adversarial Markov Decision Processes with Bandit Feedback
• Why do We Need Large Batchsizes in Contrastive Learning? A Gradient-Bias Perspective
• Globally Convergent Policy Search for Output Estimation
• To update or not to update? Neurons at equilibrium in deep models
• Grow and Merge: A Unified Framework for Continuous Categories Discovery
• OGC: Unsupervised 3D Object Segmentation from Rigid Dynamics of Point Clouds
• Learning a Condensed Frame for Memory-Efficient Video Class-Incremental Learning
• Factorized-FL: Personalized Federated Learning with Parameter Factorization & Similarity Matching
• Autoinverse: Uncertainty Aware Inversion of Neural Networks
• Bootstrapped Transformer for Offline Reinforcement Learning
• Double Check Your State Before Trusting It: Confidence-Aware Bidirectional Offline Model-Based Imagination
• Fair Wrapping for Black-box Predictions
• GT-GAN: General Purpose Time Series Synthesis with Generative Adversarial Networks
• Generic bounds on the approximation error for physics-informed (and) operator learning
• Debiased, Longitudinal and Coordinated Drug Recommendation through Multi-Visit Clinic Records
• Most Activation Functions Can Win the Lottery Without Excessive Depth
• VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training
• Truncated Matrix Power Iteration for Differentiable DAG Learning
• Robust Rent Division
• Temporally Disentangled Representation Learning
• Improving Transformer with an Admixture of Attention Heads
• Para-CFlows: $C^k$-universal diffeomorphism approximators as superior neural surrogates
• TA-GATES: An Encoding Scheme for Neural Network Architectures
• Gradient Methods Provably Converge to Non-Robust Networks
• Convolutional Neural Networks on Graphs with Chebyshev Approximation, Revisited
• Mask-based Latent Reconstruction for Reinforcement Learning
• MaskPlace: Fast Chip Placement via Reinforced Visual Representation Learning
• SwinTrack: A Simple and Strong Baseline for Transformer Tracking
• Self-supervised Amodal Video Object Segmentation
• Improving Generative Adversarial Networks via Adversarial Learning in Latent Space
• EF-BV: A Unified Theory of Error Feedback and Variance Reduction Mechanisms for Biased and Unbiased Compression in Distributed Optimization
• First is Better Than Last for Language Data Influence
• Molecule Generation by Principal Subgraph Mining and Assembling
• Conditional Independence Testing with Heteroskedastic Data and Applications to Causal Discovery
• Equivariant Graph Hierarchy-Based Neural Networks
• Semi-infinitely Constrained Markov Decision Processes
• One Positive Label is Sufficient: Single-Positive Multi-Label Learning with Label Enhancement
• Bridge the Gap Between Architecture Spaces via A Cross-Domain Predictor
• Mingling Foresight with Imagination: Model-Based Cooperative Multi-Agent Reinforcement Learning
• Top Two Algorithms Revisited
• Revisiting Graph Contrastive Learning from the Perspective of Graph Spectrum
• A Probabilistic Graph Coupling View of Dimension Reduction
• Knowledge Distillation Improves Graph Structure Augmentation for Graph Neural Networks
• LDSA: Learning Dynamic Subtask Assignment in Cooperative Multi-Agent Reinforcement Learning
• Stimulative Training of Residual Networks: A Social Psychology Perspective of Loafing
• MExMI: Pool-based Active Model Extraction Crossover Membership Inference
• S3GC: Scalable Self-Supervised Graph Clustering
• Parameter-free Dynamic Graph Embedding for Link Prediction
• Federated Submodel Optimization for Hot and Cold Data Features
• Picking on the Same Person: Does Algorithmic Monoculture lead to Outcome Homogenization?
• Causality Preserving Chaotic Transformation and Classification using Neurochaos Learning
• On Margin Maximization in Linear and ReLU Networks
• Optimal Binary Classification Beyond Accuracy
• Active Learning of Classifiers with Label and Seed Queries
• AdaptFormer: Adapting Vision Transformers for Scalable Visual Recognition
• FedPop: A Bayesian Approach for Personalised Federated Learning
• Learning to Drop Out: An Adversarial Approach to Training Sequence VAEs
• Escaping Saddle Points with Bias-Variance Reduced Local Perturbed SGD for Communication Efficient Nonconvex Distributed Learning
• Effective Adaptation in Multi-Task Co-Training for Unified Autonomous Driving
• Large Language Models are Zero-Shot Reasoners
• Descent Steps of a Relation-Aware Energy Produce Heterogeneous Graph Neural Networks
• Amplifying Membership Exposure via Data Poisoning
• Robust Graph Structure Learning via Multiple Statistical Tests
• Geometric Knowledge Distillation: Topology Compression for Graph Neural Networks
• Learning to Constrain Policy Optimization with Virtual Trust Region
• Active Surrogate Estimators: An Active Learning Approach to Label-Efficient Model Evaluation
• NodeFormer: A Scalable Graph Structure Learning Transformer for Node Classification
• Efficient Architecture Search for Diverse Tasks
• GRASP: Navigating Retrosynthetic Planning with Goal-driven Policy
• Multi-Agent Reinforcement Learning is a Sequence Modeling Problem
• Distributed Learning of Conditional Quantiles in the Reproducing Kernel Hilbert Space
• The Policy-gradient Placement and Generative Routing Neural Networks for Chip Design
• Pruning Neural Networks via Coresets and Convex Geometry: Towards No Assumptions
• Revisiting Injective Attacks on Recommender Systems
• On the Convergence of Stochastic Multi-Objective Gradient Manipulation and Beyond
• Learning to Generate Inversion-Resistant Model Explanations
• Semi-Supervised Generative Models for Multiagent Trajectories
• Unknown-Aware Domain Adversarial Learning for Open-Set Domain Adaptation
• Distributionally Robust Optimization via Ball Oracle Acceleration
• DeepMed: Semiparametric Causal Mediation Analysis with Debiased Deep Learning
• Domain Adaptation under Open Set Label Shift
• Generalization Bounds with Minimal Dependency on Hypothesis Class via Distributionally Robust Optimization
• NeMF: Neural Motion Fields for Kinematic Animation
• On Robust Multiclass Learnability
• Moment Distributionally Robust Tree Structured Prediction
• Alleviating "Posterior Collapse'' in Deep Topic Models via Policy Gradient
• Grounding Aleatoric Uncertainty for Unsupervised Environment Design
• Conditional Meta-Learning of Linear Representations
• AZ-whiteness test: a test for signal uncorrelation on spatio-temporal graphs
• Sample-Then-Optimize Batch Neural Thompson Sampling
• Efficient Adversarial Training without Attacking: Worst-Case-Aware Robust Reinforcement Learning
• Fast Bayesian Inference with Batch Bayesian Quadrature via Kernel Recombination
• EGSDE: Unpaired Image-to-Image Translation via Energy-Guided Stochastic Differential Equations
• Blessing of Depth in Linear Regression: Deeper Models Have Flatter Landscape Around the True Solution
• PKD: General Distillation Framework for Object Detectors via Pearson Correlation Coefficient
• Star Temporal Classification: Sequence Modeling with Partially Labeled Data
• Neural Stochastic Control
• Smoothed Online Convex Optimization Based on Discounted-Normal-Predictor
• Adaptive Sampling for Discovery
• Diverse Weight Averaging for Out-of-Distribution Generalization
• Counterfactual Temporal Point Processes
• Sparse Winning Tickets are Data-Efficient Image Recognizers
• Improved Regret Analysis for Variance-Adaptive Linear Bandits and Horizon-Free Linear Mixture MDPs
• Reduction Algorithms for Persistence Diagrams of Networks: CoralTDA and PrunIT
• Approximation with CNNs in Sobolev Space: with Applications to Classification
• Tracking Functional Changes in Nonstationary Signals with Evolutionary Ensemble Bayesian Model for Robust Neural Decoding
• A Unified Convergence Theorem for Stochastic Optimization Methods
• On Embeddings for Numerical Features in Tabular Deep Learning
• Near-Optimal Collaborative Learning in Bandits
• Increasing the Scope as You Learn: Adaptive Bayesian Optimization in Nested Subspaces
• Iron: Private Inference on Transformers
• Towards Disentangling Information Paths with Coded ResNeXt
• Adaptive Multi-stage Density Ratio Estimation for Learning Latent Space Energy-based Model
• Flamingo: a Visual Language Model for Few-Shot Learning
• Learning to Re-weight Examples with Optimal Transport for Imbalanced Classification
• ZooD: Exploiting Model Zoo for Out-of-Distribution Generalization
• Torsional Diffusion for Molecular Conformer Generation
• Beyond Time-Average Convergence: Near-Optimal Uncoupled Online Learning via Clairvoyant Multiplicative Weights Update
• Approximate Euclidean lengths and distances beyond Johnson-Lindenstrauss
• A consistently adaptive trust-region method
• Order-Invariant Cardinality Estimators Are Differentially Private
• Spectral Bias in Practice: The Role of Function Frequency in Generalization
• Task-level Differentially Private Meta Learning
• Distributed Inverse Constrained Reinforcement Learning for Multi-agent Systems
• WaveBound: Dynamic Error Bounds for Stable Time Series Forecasting
• Self-Supervised Learning of Brain Dynamics from Broad Neuroimaging Data
• Gradient flow dynamics of shallow ReLU networks for square loss and orthogonal inputs
• Ensemble of Averages: Improving Model Selection and Boosting Performance in Domain Generalization
• Log-Polar Space Convolution Layers
• Efficient Training of Low-Curvature Neural Networks
• Self-Supervised Fair Representation Learning without Demographics
• Nonlinear MCMC for Bayesian Machine Learning
• Scale-invariant Learning by Physics Inversion
• On Non-Linear operators for Geometric Deep Learning
• A Geometric Perspective on Variational Autoencoders
• Contrastive Graph Structure Learning via Information Bottleneck for Recommendation
• Iterative Structural Inference of Directed Graphs
• PDSketch: Integrated Domain Programming, Learning, and Planning
• Off-Policy Evaluation with Policy-Dependent Optimization Response
• Interpolation and Regularization for Causal Learning
• Confidence-based Reliable Learning under Dual Noises
• Dynamic Inverse Reinforcement Learning for Characterizing Animal Behavior
• DReS-FL: Dropout-Resilient Secure Federated Learning for Non-IID Clients via Secret Data Sharing
• Black-Box Generalization: Stability of Zeroth-Order Learning
• Label-Aware Global Consistency for Multi-Label Learning with Single Positive Labels
• Emergent Communication: Generalization and Overfitting in Lewis Games
• Latent Planning via Expansive Tree Search
• Near-Optimal Regret for Adversarial MDP with Delayed Bandit Feedback
• RTFormer: Efficient Design for Real-Time Semantic Segmentation with Transformer
• The Phenomenon of Policy Churn
• Optimal-er Auctions through Attention
• Sampling with Riemannian Hamiltonian Monte Carlo in a Constrained Space
• Defending Against Adversarial Attacks via Neural Dynamic System
• Association Graph Learning for Multi-Task Classification with Category Shifts
• Weakly Supervised Representation Learning with Sparse Perturbations
• Chain-of-Thought Prompting Elicits Reasoning in Large Language Models
• Learning Superpoint Graph Cut for 3D Instance Segmentation
• First Contact: Unsupervised Human-Machine Co-Adaptation via Mutual Information Maximization
• CyCLIP: Cyclic Contrastive Language-Image Pretraining
• Amortized Inference for Heterogeneous Reconstruction in Cryo-EM
• Neural Stochastic PDEs: Resolution-Invariant Learning of Continuous Spatiotemporal Dynamics
• Sobolev Acceleration and Statistical Optimality for Learning Elliptic Equations via Gradient Descent
• Listen to Interpret: Post-hoc Interpretability for Audio Networks with NMF
• An $\alpha$-No-Regret Algorithm For Graphical Bilinear Bandits
• Perfect Sampling from Pairwise Comparisons
• Value Function Decomposition for Iterative Design of Reinforcement Learning Agents
• Constraining Gaussian Processes to Systems of Linear Ordinary Differential Equations
• VAEL: Bridging Variational Autoencoders and Probabilistic Logic Programming
• Conformal Off-Policy Prediction in Contextual Bandits
• Constrained Update Projection Approach to Safe Policy Optimization
• Global Linear and Local Superlinear Convergence of IRLS for Non-Smooth Robust Regression
• A Fourier Approach to Mixture Learning
• LECO: Learnable Episodic Count for Task-Specific Intrinsic Reward
• Domain Generalization without Excess Empirical Risk
• Navigating Memory Construction by Global Pseudo-Task Simulation for Continual Learning
• Optimal Transport of Classifiers to Fairness
• FedSR: A Simple and Effective Domain Generalization Method for Federated Learning
• Using Partial Monotonicity in Submodular Maximization
• When Do Flat Minima Optimizers Work?
• Revisiting Non-Parametric Matching Cost Volumes for Robust and Generalizable Stereo Matching
• Large-scale Optimization of Partial AUC in a Range of False Positive Rates
• Learning in Congestion Games with Bandit Feedback
• TreeMoCo: Contrastive Neuron Morphology Representation Learning
• Near-Optimal Sample Complexity Bounds for Constrained MDPs
• Fairness Transferability Subject to Bounded Distribution Shift
• The Burer-Monteiro SDP method can fail even above the Barvinok-Pataki bound
• WeightedSHAP: analyzing and improving Shapley based feature attributions
• How to talk so AI will learn: Instructions, descriptions, and autonomy
• Improved Algorithms for Neural Active Learning
• Global Convergence of Direct Policy Search for State-Feedback $\mathcal{H}_\infty$ Robust Control: A Revisit of Nonsmooth Synthesis with Goldstein Subdifferential
• Nonlinear Sufficient Dimension Reduction with a Stochastic Neural Network
• Bayesian inference via sparse Hamiltonian flows
• On Batch Teaching with Sample Complexity Bounded by VCD
• AVLEN: Audio-Visual-Language Embodied Navigation in 3D Environments
• Model-based Lifelong Reinforcement Learning with Bayesian Exploration
• projUNN: efficient method for training deep networks with unitary matrices
• Staggered Rollout Designs Enable Causal Inference Under Interference Without Network Knowledge
• KERPLE: Kernelized Relative Positional Embedding for Length Extrapolation
• An Information-Theoretic Framework for Deep Learning
• ORIENT: Submodular Mutual Information Measures for Data Subset Selection under Distribution Shift
• Insights into Pre-training via Simpler Synthetic Tasks
• Attracting and Dispersing: A Simple Approach for Source-free Domain Adaptation
• Nearly Optimal Algorithms for Linear Contextual Bandits with Adversarial Corruptions
• Density-driven Regularization for Out-of-distribution Detection
• Rapid Model Architecture Adaption for Meta-Learning
• Finding Correlated Equilibrium of Constrained Markov Game: A Primal-Dual Approach
• Hyperbolic Embedding Inference for Structured Multi-Label Prediction
• AutoMTL: A Programming Framework for Automating Efficient Multi-Task Learning
• Wavelet Feature Maps Compression for Image-to-Image CNNs
• CoPur: Certifiably Robust Collaborative Inference via Feature Purification
• Interventions, Where and How? Experimental Design for Causal Models at Scale
• Efficient Non-Parametric Optimizer Search for Diverse Tasks
• Seeing the forest and the tree: Building representations of both individual and collective dynamics with transformers
• Deep Architecture Connectivity Matters for Its Convergence: A Fine-Grained Analysis
• Scaling Multimodal Pre-Training via Cross-Modality Gradient Harmonization
• A Character-Level Length-Control Algorithm for Non-Autoregressive Sentence Summarization
• The Privacy Onion Effect: Memorization is Relative
• Recursive Reasoning in Minimax Games: A Level $k$ Gradient Play Method
• Deep Ensembles Work, But Are They Necessary?
• Tight Lower Bounds on Worst-Case Guarantees for Zero-Shot Learning with Attributes
• Variational Model Perturbation for Source-Free Domain Adaptation
• Generative multitask learning mitigates target-causing confounding
• Draft-and-Revise: Effective Image Generation with Contextual RQ-Transformer
• Acceleration in Distributed Sparse Regression
• Learning Two-Player Markov Games: Neural Function Approximation and Correlated Equilibrium
• Repairing Neural Networks by Leaving the Right Past Behind
• [Re] Privacy-preserving collaborative learning with automatic transformation search
• Sequence Model Imitation Learning with Unobserved Contexts
• GULP: a prediction-based metric between representations
• Efficient Frameworks for Generalized Low-Rank Matrix Bandit Problems
• Composition Theorems for Interactive Differential Privacy
• On the Global Convergence Rates of Decentralized Softmax Gradient Play in Markov Potential Games
• Robust Generalized Method of Moments: A Finite Sample Viewpoint
• Boosting Barely Robust Learners: A New Perspective on Adversarial Robustness
• Regret Bounds for Risk-Sensitive Reinforcement Learning
• Semi-supervised Active Linear Regression
• Near-Isometric Properties of Kronecker-Structured Random Tensor Embeddings
• Riemannian Diffusion Models
• Towards Safe Reinforcement Learning with a Safety Editor Policy
• On the Safety of Interpretable Machine Learning: A Maximum Deviation Approach
• The Implicit Delta Method
• Bellman Residual Orthogonalization for Offline Reinforcement Learning
• Meta-Learning Dynamics Forecasting Using Task Inference
• On Scrambling Phenomena for Randomly Initialized Recurrent Networks
• Data-Efficient Augmentation for Training Neural Networks
• Beyond black box densities: Parameter learning for the deviated components
• Robust Learning against Relational Adversaries
• Policy Optimization for Markov Games: Unified Framework and Faster Convergence
• Continuously Tempered PDMP samplers
• Uncalibrated Models Can Improve Human-AI Collaboration
• Few-Shot Non-Parametric Learning with Deep Latent Variable Model
• Emergent Graphical Conventions in a Visual Communication Game
• Chain of Thought Imitation with Procedure Cloning
• Conformalized Fairness via Quantile Regression
• Improving Self-Supervised Learning by Characterizing Idealized Representations
• Learning Options via Compression
• Rapidly Mixing Multiple-try Metropolis Algorithms for Model Selection Problems
• Understanding Hyperdimensional Computing for Parallel Single-Pass Learning
• Functional Indirection Neural Estimator for Better Out-of-distribution Generalization
• Few-shot Learning for Feature Selection with Hilbert-Schmidt Independence Criterion
• Bayesian Spline Learning for Equation Discovery of Nonlinear Dynamics with Quantified Uncertainty
• Sketch-GNN: Scalable Graph Neural Networks with Sublinear Training Complexity
• Active Learning Polynomial Threshold Functions
• Efficient and Near-Optimal Smoothed Online Learning for Generalized Linear Functions
• Signal Propagation in Transformers: Theoretical Perspectives and the Role of Rank Collapse
• An Analysis of Ensemble Sampling
• Towards Understanding the Condensation of Neural Networks at Initial Training
• Fair Infinitesimal Jackknife: Mitigating the Influence of Biased Training Data Points Without Refitting
• Optimal Scaling for Locally Balanced Proposals in Discrete Spaces
• End-to-end Algorithm Synthesis with Recurrent Networks: Extrapolation without Overthinking
• Domain Adaptation meets Individual Fairness. And they get along.
• Free Probability for predicting the performance of feed-forward fully connected neural networks
• Conformal Prediction with Temporal Quantile Adjustments
• Using natural language and program abstractions to instill human inductive biases in machines
• Generalizing Goal-Conditioned Reinforcement Learning with Variational Causal Reasoning
• Neurosymbolic Deep Generative Models for Sequence Data with Relational Constraints
• Polynomial time guarantees for the Burer-Monteiro method
• Scalable design of Error-Correcting Output Codes using Discrete Optimization with Graph Coloring
• On Deep Generative Models for Approximation and Estimation of Distributions on Manifolds
• Nest Your Adaptive Algorithm for Parameter-Agnostic Nonconvex Minimax Optimization
• Simple Unsupervised Object-Centric Learning for Complex and Naturalistic Videos
• Revisiting Optimal Convergence Rate for Smooth and Non-convex Stochastic Decentralized Optimization
• Physics-Informed Implicit Representations of Equilibrium Network Flows
• Learning Generalized Policy Automata for Relational Stochastic Shortest Path Problems
• Simplified Graph Convolution with Heterophily
• DMAP: a Distributed Morphological Attention Policy for learning to locomote with a changing body
• Byzantine-tolerant federated Gaussian process regression for streaming data
• Distributionally Adaptive Meta Reinforcement Learning
• Submodular Maximization in Clean Linear Time
• Amortized Proximal Optimization
• On Learning Fairness and Accuracy on Multiple Subgroups
• HUMUS-Net: Hybrid Unrolled Multi-scale Network Architecture for Accelerated MRI Reconstruction
• On the Symmetries of Deep Learning Models and their Internal Representations
• Calibrated Data-Dependent Constraints with Exact Satisfaction Guarantees
• Second Thoughts are Best: Learning to Re-Align With Human Values from Text Edits
• Decision-based Black-box Attack Against Vision Transformers via Patch-wise Adversarial Removal
• FourierFormer: Transformer Meets Generalized Fourier Integral Theorem
• In What Ways Are Deep Neural Networks Invariant and How Should We Measure This?
• Faster and Scalable Algorithms for Densest Subgraph and Decomposition
• Co-Modality Graph Contrastive Learning for Imbalanced Node Classification
• ACIL: Analytic Class-Incremental Learning with Absolute Memorization and Privacy Protection
• Zeroth-Order Hard-Thresholding: Gradient Error vs. Expansivity
• Non-Linguistic Supervision for Contrastive Learning of Sentence Embeddings
• A Non-Asymptotic Moreau Envelope Theory for High-Dimensional Generalized Linear Models
• A General Framework for Auditing Differentially Private Machine Learning
• Minimax Optimal Algorithms for Fixed-Budget Best Arm Identification
• Pruning’s Effect on Generalization Through the Lens of Training and Regularization
• Kernel similarity matching with Hebbian networks
• QC-StyleGAN - Quality Controllable Image Generation and Manipulation
• Human-AI Shared Control via Policy Dissection
• Label-invariant Augmentation for Semi-Supervised Graph Classification
• Preservation of the Global Knowledge by Not-True Distillation in Federated Learning
• Using Embeddings for Causal Estimation of Peer Influence in Social Networks
• A Unifying Framework of Off-Policy General Value Function Evaluation
• TaSIL: Taylor Series Imitation Learning
• Asymptotic Behaviors of Projected Stochastic Approximation: A Jump Diffusion Perspective
• VF-PS: How to Select Important Participants in Vertical Federated Learning, Efficiently and Securely?
• LISA: Learning Interpretable Skill Abstractions from Language
• NSNet: A General Neural Probabilistic Framework for Satisfiability Problems
• Model Preserving Compression for Neural Networks
• Effects of Data Geometry in Early Deep Learning
• Minimax-Optimal Multi-Agent RL in Markov Games With a Generative Model
• Tight Mutual Information Estimation With Contrastive Fenchel-Legendre Optimization
• Pre-Trained Model Reusability Evaluation for Small-Data Transfer Learning
• Graph Few-shot Learning with Task-specific Structures
• Offline Goal-Conditioned Reinforcement Learning via $f$-Advantage Regression
• Old can be Gold: Better Gradient Flow can Make Vanilla-GCNs Great Again
• Faster Deep Reinforcement Learning with Slower Online Network
• An Asymptotically Optimal Batched Algorithm for the Dueling Bandit Problem
• Beyond Not-Forgetting: Continual Learning with Backward Knowledge Transfer
• Distributed Distributionally Robust Optimization with Non-Convex Objectives
• Continuous Deep Q-Learning in Optimal Control Problems: Normalized Advantage Functions Analysis
• Merging Models with Fisher-Weighted Averaging
• Path Independent Equilibrium Models Can Better Exploit Test-Time Computation
• Private Graph All-Pairwise-Shortest-Path Distance Release with Improved Error Rate
• A Theory of PAC Learnability under Transformation Invariances
• Global Convergence of Federated Learning for Mixed Regression
• Segmenting Moving Objects via an Object-Centric Layered Representation
• Invariance Learning based on Label Hierarchy
• Online Algorithms for the Santa Claus Problem
• Federated Learning from Pre-Trained Models: A Contrastive Learning Approach
• When are Offline Two-Player Zero-Sum Markov Games Solvable?
• Tree Mover's Distance: Bridging Graph Metrics and Stability of Graph Neural Networks
• Provably Efficient Offline Multi-agent Reinforcement Learning via Strategy-wise Bonus
• Implicit Bias of Gradient Descent on Reparametrized Models: On Equivalence to Mirror Descent
• Few-Shot Audio-Visual Learning of Environment Acoustics
• Redundancy-Free Message Passing for Graph Neural Networks
• SemMAE: Semantic-Guided Masking for Learning Masked Autoencoders
• Weighted Distillation with Unlabeled Examples
• Mixture-of-Experts with Expert Choice Routing
• The Stability-Efficiency Dilemma: Investigating Sequence Length Warmup for Training GPT Models
• Diffusion-LM Improves Controllable Text Generation
• Self-Supervised Pretraining for Large-Scale Point Clouds
• Invariant and Transportable Representations for Anti-Causal Domain Shifts
• Sparsity in Continuous-Depth Neural Networks
• A Variational Edge Partition Model for Supervised Graph Representation Learning
• A Simple Approach to Automated Spectral Clustering
• Point Transformer V2: Grouped Vector Attention and Partition-based Pooling
• Accelerated Training of Physics-Informed Neural Networks (PINNs) using Meshless Discretizations
• Fault-Aware Neural Code Rankers
• PAC-Bayes Compression Bounds So Tight That They Can Explain Generalization
• A simple but strong baseline for online continual learning: Repeated Augmented Rehearsal
• Off-Policy Evaluation for Episodic Partially Observable Markov Decision Processes under Non-Parametric Models
• Learning Symmetric Rules with SATNet
• Consistent Interpolating Ensembles via the Manifold-Hilbert Kernel
• Get More at Once: Alternating Sparse Training with Gradient Correction
• Learning Fractional White Noises in Neural Stochastic Differential Equations
• “Why Not Other Classes?”: Towards Class-Contrastive Back-Propagation Explanations
• Why Robust Generalization in Deep Learning is Difficult: Perspective of Expressive Power
• Training with More Confidence: Mitigating Injected and Natural Backdoors During Training
• Iterative Feature Matching: Toward Provable Domain Generalization with Logarithmic Environments
• Batch Multi-Fidelity Active Learning with Budget Constraints
• Between Stochastic and Adversarial Online Convex Optimization: Improved Regret Bounds via Smoothness
• Time-Conditioned Dances with Simplicial Complexes: Zigzag Filtration Curve based Supra-Hodge Convolution Networks for Time-series Forecasting
• Constrained Stochastic Nonconvex Optimization with State-dependent Markov Data
• Integral Probability Metrics PAC-Bayes Bounds
• Skills Regularized Task Decomposition for Multi-task Offline Reinforcement Learning
• M2N: Mesh Movement Networks for PDE Solvers
• Exploiting the Relationship Between Kendall's Rank Correlation and Cosine Similarity for Attribution Protection
• Understanding and Improving Robustness of Vision Transformers through Patch-based Negative Augmentation
• Gaussian Copula Embeddings
• Transferring Pre-trained Multimodal Representations with Cross-modal Similarity Matching
• CoNSoLe: Convex Neural Symbolic Learning
• Maximum-Likelihood Inverse Reinforcement Learning with Finite-Time Guarantees
• Meta-Auto-Decoder for Solving Parametric Partial Differential Equations
• Non-Stationary Bandits under Recharging Payoffs: Improved Planning with Sublinear Regret
• Long-Form Video-Language Pre-Training with Multimodal Temporal Contrastive Learning
• PlasticityNet: Learning to Simulate Metal, Sand, and Snow for Optimization Time Integration
• [Re] Value Alignment Verification
• Nearly-Tight Bounds for Testing Histogram Distributions
• Rethinking and Scaling Up Graph Contrastive Learning: An Extremely Efficient Approach with Group Discrimination
• Uncertainty Estimation for Multi-view Data: The Power of Seeing the Whole Picture
• Reinforcement Learning with Automated Auxiliary Loss Search
• Tractable Function-Space Variational Inference in Bayesian Neural Networks
• Are all Frames Equal? Active Sparse Labeling for Video Action Detection
• Unsupervised Learning under Latent Label Shift
• You Can’t Count on Luck: Why Decision Transformers and RvS Fail in Stochastic Environments
• Provable Subspace Identification Under Post-Nonlinear Mixtures
• Truly Deterministic Policy Optimization
• Active Learning Helps Pretrained Models Learn the Intended Task
• A Consolidated Cross-Validation Algorithm for Support Vector Machines via Data Reduction
• Giving Feedback on Interactive Student Programs with Meta-Exploration
• On Leave-One-Out Conditional Mutual Information For Generalization
• High-dimensional Asymptotics of Feature Learning: How One Gradient Step Improves the Representation
• Global Optimal K-Medoids Clustering of One Million Samples
• A Scalable Deterministic Global Optimization Algorithm for Training Optimal Decision Tree
• What Can Transformers Learn In-Context? A Case Study of Simple Function Classes
• GALOIS: Boosting Deep Reinforcement Learning via Generalizable Logic Synthesis
• Provably Feedback-Efficient Reinforcement Learning via Active Reward Learning
• Data-Driven Offline Decision-Making via Invariant Representation Learning
• When Does Group Invariant Learning Survive Spurious Correlations?
• On Elimination Strategies for Bandit Fixed-Confidence Identification
• So3krates: Equivariant attention for interactions on arbitrary length-scales in molecular systems
• Making Look-Ahead Active Learning Strategies Feasible with Neural Tangent Kernels
• DHRL: A Graph-Based Approach for Long-Horizon and Sparse Hierarchical Reinforcement Learning
• Improving Diffusion Models for Inverse Problems using Manifold Constraints
• DARE: Disentanglement-Augmented Rationale Extraction
• Symmetry-induced Disentanglement on Graphs
• Learning in Observable POMDPs, without Computationally Intractable Oracles
• When to Ask for Help: Proactive Interventions in Autonomous Reinforcement Learning
• Spherization Layer: Representation Using Only Angles
• Grounded Reinforcement Learning: Learning to Win the Game under Human Commands
• How Powerful are K-hop Message Passing Graph Neural Networks
• MEMO: Test Time Robustness via Adaptation and Augmentation
• Redundant representations help generalization in wide neural networks
• Dynamic Learning in Large Matching Markets
• Near-Optimal Goal-Oriented Reinforcement Learning in Non-Stationary Environments
• Towards Understanding the Mixture-of-Experts Layer in Deep Learning
• A time-resolved theory of information encoding in recurrent neural networks
• Coresets for Relational Data and The Applications
• Coresets for Wasserstein Distributionally Robust Optimization Problems
• Lazy and Fast Greedy MAP Inference for Determinantal Point Process
• FlowHMM: Flow-based continuous hidden Markov models
• Max-Min Off-Policy Actor-Critic Method Focusing on Worst-Case Robustness to Model Misspecification
• An Adaptive Deep RL Method for Non-Stationary Environments with Piecewise Stable Context
• Experimental Design for Linear Functionals in Reproducing Kernel Hilbert Spaces
• Imbalance Trouble: Revisiting Neural-Collapse Geometry
• Injecting Domain Knowledge from Empirical Interatomic Potentials to Neural Networks for Predicting Material Properties
• How Mask Matters: Towards Theoretical Understandings of Masked Autoencoders
• WT-MVSNet: Window-based Transformers for Multi-view Stereo
• Models Out of Line: A Fourier Lens on Distribution Shift Robustness
• SCINet: Time Series Modeling and Forecasting with Sample Convolution and Interaction
• Chromatic Correlation Clustering, Revisited
• A Reduction to Binary Approach for Debiasing Multiclass Datasets
• MetricFormer: A Unified Perspective of Correlation Exploring in Similarity Learning
• Asynchronous SGD Beats Minibatch SGD Under Arbitrary Delays
• Revisiting Neural Scaling Laws in Language and Vision
• Towards Consistency in Adversarial Classification
• Last-Iterate Convergence of Optimistic Gradient Method for Monotone Variational Inequalities
• Graph Convolution Network based Recommender Systems: Learning Guarantee and Item Mixture Powered Strategy
• Revisit last-iterate convergence of mSGD under milder requirement on step size
• Non-Monotonic Latent Alignments for CTC-Based Non-Autoregressive Machine Translation
• Joint Learning of 2D-3D Weakly Supervised Semantic Segmentation
• Graph Coloring via Neural Networks for Haplotype Assembly and Viral Quasispecies Reconstruction
• Optimal Positive Generation via Latent Transformation for Contrastive Learning
• Neural-Symbolic Entangled Framework for Complex Query Answering
• Multiagent Q-learning with Sub-Team Coordination
• Sound and Complete Verification of Polynomial Networks
• Laplacian Autoencoders for Learning Stochastic Representations
• Oracle Inequalities for Model Selection in Offline Reinforcement Learning
• Revisiting Active Sets for Gaussian Process Decoders
• Bounding and Approximating Intersectional Fairness through Marginal Fairness
• MAtt: A Manifold Attention Network for EEG Decoding
• BinauralGrad: A Two-Stage Conditional Diffusion Probabilistic Model for Binaural Audio Synthesis
• Collaborative Decision Making Using Action Suggestions
• Dynamics of SGD with Stochastic Polyak Stepsizes: Truly Adaptive Variants and Convergence to Exact Solution
• A gradient estimator via L1-randomization for online zero-order optimization with two point feedback
• Bring Your Own Algorithm for Optimal Differentially Private Stochastic Minimax Optimization
• Selective compression learning of latent representations for variable-rate image compression
• Neural Network Architecture Beyond Width and Depth
• On the relationship between variational inference and auto-associative memory
• Sparse Probabilistic Circuits via Pruning and Growing
• When to Intervene: Learning Optimal Intervention Policies for Critical Events
• Smoothed Embeddings for Certified Few-Shot Learning
• An Analytical Theory of Curriculum Learning in Teacher-Student Networks
• Black-box coreset variational inference
• Distilling Representations from GAN Generator via Squeeze and Span
• Generalization Analysis of Message Passing Neural Networks on Large Random Graphs
• Meta-Learning with Self-Improving Momentum Target
• Reinforcement Learning in a Birth and Death Process: Breaking the Dependence on the State Space
• Sequence-to-Set Generative Models
• What Makes Graph Neural Networks Miscalibrated?
• A Win-win Deal: Towards Sparse and Robust Pre-trained Language Models
• Consistency of Constrained Spectral Clustering under Graph Induced Fair Planted Partitions
• A Regret-Variance Trade-Off in Online Learning
• Optimal Brain Compression: A Framework for Accurate Post-Training Quantization and Pruning
• Not too little, not too much: a theoretical analysis of graph (over)smoothing
• On A Mallows-type Model For (Ranked) Choices
• Inverse Design for Fluid-Structure Interactions using Graph Network Simulators
• Towards a Standardised Performance Evaluation Protocol for Cooperative MARL
• On the Learning Mechanisms in Physical Reasoning
• Joint Entropy Search For Maximally-Informed Bayesian Optimization
• Benign Overfitting in Two-layer Convolutional Neural Networks
• Proximal Point Imitation Learning
• On the Robustness of Graph Neural Diffusion to Topology Perturbations
• Power and limitations of single-qubit native quantum neural networks
• A Characterization of Semi-Supervised Adversarially Robust PAC Learnability
• Accelerating SGD for Highly Ill-Conditioned Huge-Scale Online Matrix Completion
• Accelerated Primal-Dual Gradient Method for Smooth and Convex-Concave Saddle-Point Problems with Bilinear Coupling
• Distilled Gradient Aggregation: Purify Features for Input Attribution in the Deep Neural Network
• Sequential Information Design: Learning to Persuade in the Dark
• Optimal Weak to Strong Learning
• Unsupervised Learning of Group Invariant and Equivariant Representations
• On the Approximation of Cooperative Heterogeneous Multi-Agent Reinforcement Learning (MARL) using Mean Field Control (MFC)
• Estimating the Arc Length of the Optimal ROC Curve and Lower Bounding the Maximal AUC
• Disentangling Causal Effects from Sets of Interventions in the Presence of Unobserved Confounders
• A Reparametrization-Invariant Sharpness Measure Based on Information Geometry
• Bayesian Active Learning with Fully Bayesian Gaussian Processes
• Log-Concave and Multivariate Canonical Noise Distributions for Differential Privacy
• On Measuring Excess Capacity in Neural Networks
• General Cutting Planes for Bound-Propagation-Based Neural Network Verification
• Unsupervised Adaptation from Repeated Traversals for Autonomous Driving
• Fine-Grained Semantically Aligned Vision-Language Pre-Training
• On Sample Optimality in Personalized Collaborative and Federated Learning
• Using Mixup as a Regularizer Can Surprisingly Improve Accuracy & Out-of-Distribution Robustness
• Sharper Convergence Guarantees for Asynchronous SGD for Distributed and Federated Learning
• A Variant of Anderson Mixing with Minimal Memory Size
• Augmented RBMLE-UCB Approach for Adaptive Control of Linear Quadratic Systems
• Improved techniques for deterministic l2 robustness
• Real-Valued Backpropagation is Unsuitable for Complex-Valued Neural Networks
• Anonymized Histograms in Intermediate Privacy Models
• Relaxing Equivariance Constraints with Non-stationary Continuous Filters
• MCL-GAN: Generative Adversarial Networks with Multiple Specialized Discriminators
• Sparse Gaussian Process Hyperparameters: Optimize or Integrate?
• Sleeper Agent: Scalable Hidden Trigger Backdoors for Neural Networks Trained from Scratch
• Learning to Branch with Tree MDPs
• Fine-tuning Language Models over Slow Networks using Activation Quantization with Guarantees
• Probing Classifiers are Unreliable for Concept Removal and Detection
• Graph Learning Assisted Multi-Objective Integer Programming
• Randomized Sketches for Clustering: Fast and Optimal Kernel $k$-Means
• Escaping Saddle Points for Effective Generalization on Class-Imbalanced Data
• Certifying Robust Graph Classification under Orthogonal Gromov-Wasserstein Threats
• On the Representation Collapse of Sparse Mixture of Experts
• Information bottleneck theory of high-dimensional regression: relevancy, efficiency and optimality
• Statistically Meaningful Approximation: a Case Study on Approximating Turing Machines with Transformers
• A Data-Augmentation Is Worth A Thousand Samples: Analytical Moments And Sampling-Free Training
• Partial Identification of Treatment Effects with Implicit Generative Models
• Learning Neural Acoustic Fields
• Variance Reduced ProxSkip: Algorithm, Theory and Application to Federated Learning
• Local Metric Learning for Off-Policy Evaluation in Contextual Bandits with Continuous Actions
• A contrastive rule for meta-learning
• Meta Reinforcement Learning with Finite Training Tasks - a Density Estimation Approach
• A gradient sampling method with complexity guarantees for Lipschitz functions in high and low dimensions
• Regularized Molecular Conformation Fields
• You Never Stop Dancing: Non-freezing Dance Generation via Bank-constrained Manifold Projection
• Risk-Driven Design of Perception Systems
• Langevin Autoencoders for Learning Deep Latent Variable Models
• Neural Estimation of Submodular Functions with Applications to Differentiable Subset Selection
• Learning on Arbitrary Graph Topologies via Predictive Coding
• Multi-Lingual Acquisition on Multimodal Pre-training for Cross-modal Retrieval
• Semantic Exploration from Language Abstractions and Pretrained Representations
• A Unified Sequence Interface for Vision Tasks
• Is Integer Arithmetic Enough for Deep Learning Training?
• Confident Adaptive Language Modeling
• Learning Dynamical Systems via Koopman Operator Regression in Reproducing Kernel Hilbert Spaces
• 3D Concept Grounding on Neural Fields
• A Solver-free Framework for Scalable Learning in Neural ILP Architectures
• Unsupervised Object Representation Learning using Translation and Rotation Group Equivariant VAE
• Beyond Rewards: a Hierarchical Perspective on Offline Multiagent Behavioral Analysis
• Luckiness in Multiscale Online Learning
• Effective Dimension in Bandit Problems under Censorship
• In Defense of the Unitary Scalarization for Deep Multi-Task Learning
• Beyond IID: data-driven decision-making in heterogeneous environments
• Scalable Multi-agent Covering Option Discovery based on Kronecker Graphs
• Private Multiparty Perception for Navigation
• Group Meritocratic Fairness in Linear Contextual Bandits
• Deep Equilibrium Approaches to Diffusion Models
• Addressing Leakage in Concept Bottleneck Models
• Evolution of Neural Tangent Kernels under Benign and Adversarial Training
• The least-control principle for local learning at equilibrium
• Where2comm: Communication-Efficient Collaborative Perception via Spatial Confidence Maps
• Thor: Wielding Hammers to Integrate Language Models and Automated Theorem Provers
• PhysGNN: A Physics--Driven Graph Neural Network Based Model for Predicting Soft Tissue Deformation in Image--Guided Neurosurgery
• Archimedes Meets Privacy: On Privately Estimating Quantiles in High Dimensions Under Minimal Assumptions
• Better SGD using Second-order Momentum
• Learning from Few Samples: Transformation-Invariant SVMs with Composition and Locality at Multiple Scales
• DevFly: Bio-Inspired Development of Binary Connections for Locality Preserving Sparse Codes
• Multi-agent Dynamic Algorithm Configuration
• Predictive Coding beyond Gaussian Distributions
• Jump Self-attention: Capturing High-order Statistics in Transformers
• Invariance Learning in Deep Neural Networks with Differentiable Laplace Approximations
• RISE: Robust Individualized Decision Learning with Sensitive Variables
• Efficient and Stable Fully Dynamic Facility Location
• Envy-free Policy Teaching to Multiple Agents
• VaiPhy: a Variational Inference Based Algorithm for Phylogeny
• Active Learning with Safety Constraints
• Trustworthy Monte Carlo
• Learning-Augmented Algorithms for Online Linear and Semidefinite Programming
• Near-Optimal Correlation Clustering with Privacy
• Neural Attentive Circuits
• Intra-agent speech permits zero-shot task acquisition
• MACK: Multimodal Aligned Conceptual Knowledge for Unpaired Image-text Matching
• Robustness to Label Noise Depends on the Shape of the Noise Distribution
• A Theoretical Study on Solving Continual Learning
• Anytime-Valid Inference For Multinomial Count Data
• Scalable and Efficient Non-adaptive Deterministic Group Testing
• Hierarchical Agglomerative Graph Clustering in Poly-Logarithmic Depth
• Variable-rate hierarchical CPC leads to acoustic unit discovery in speech
• SoteriaFL: A Unified Framework for Private Federated Learning with Communication Compression
• Contextual Dynamic Pricing with Unknown Noise: Explore-then-UCB Strategy and Improved Regrets
• Distributed Online Convex Optimization with Compressed Communication
• GlanceNets: Interpretable, Leak-proof Concept-based Models
• BEER: Fast $O(1/T)$ Rate for Decentralized Nonconvex Optimization with Communication Compression
• On the Effectiveness of Persistent Homology
• The Effects of Regularization and Data Augmentation are Class Dependent
• On the Stability and Scalability of Node Perturbation Learning
• Trimmed Maximum Likelihood Estimation for Robust Generalized Linear Model
• Benefits of Additive Noise in Composing Classes with Bounded Capacity
• EZNAS: Evolving Zero-Cost Proxies For Neural Architecture Scoring
• Proppo: a Message Passing Framework for Customizable and Composable Learning Algorithms
• Towards a Unified Framework for Uncertainty-aware Nonlinear Variable Selection with Theoretical Guarantees
• Tempo: Accelerating Transformer-Based Model Training through Memory Footprint Reduction
• CS-Shapley: Class-wise Shapley Values for Data Valuation in Classification
• A New Family of Generalization Bounds Using Samplewise Evaluated CMI
• Learning to Reconstruct Missing Data from Spatiotemporal Graphs with Sparse Observations
• On the Adversarial Robustness of Mixture of Experts
• Graph Neural Networks are Dynamic Programmers
• K-LITE: Learning Transferable Visual Models with External Knowledge
• Mesoscopic modeling of hidden spiking neurons
• Self-Supervised Learning Through Efference Copies
• Self-Explaining Deviations for Coordination
• Multi-Objective Deep Learning with Adaptive Reference Vectors
• Overparameterization from Computational Constraints
• AUTOMATA: Gradient Based Data Subset Selection for Compute-Efficient Hyper-parameter Tuning
• Anchor-Changing Regularized Natural Policy Gradient for Multi-Objective Reinforcement Learning
• On the Generalization Power of the Overfitted Three-Layer Neural Tangent Kernel Model
• Provably Adversarially Robust Detection of Out-of-Distribution Data (Almost) for Free
• Modular Flows: Differential Molecular Generation
• Bridging Central and Local Differential Privacy in Data Acquisition Mechanisms
• PAC Prediction Sets for Meta-Learning
• Diffusion Models as Plug-and-Play Priors
• MorphTE: Injecting Morphology in Tensorized Embeddings
• Trajectory balance: Improved credit assignment in GFlowNets
• On Convergence of FedProx: Local Dissimilarity Invariant Bounds, Non-smoothness and Beyond
• Task-Free Continual Learning via Online Discrepancy Distance Learning
• Improved Differential Privacy for SGD via Optimal Private Linear Operators on Adaptive Streams
• Evaluation beyond Task Performance: Analyzing Concepts in AlphaZero in Hex
• Benchopt: Reproducible, efficient and collaborative optimization benchmarks
• RNNs of RNNs: Recursive Construction of Stable Assemblies of Recurrent Neural Networks
• Nonparametric Uncertainty Quantification for Single Deterministic Neural Network
• Controlled Sparsity via Constrained Optimization or: How I Learned to Stop Tuning Penalties and Love Constraints
• Discovering and Overcoming Limitations of Noise-engineered Data-free Knowledge Distillation
• Object Representations as Fixed Points: Training Iterative Refinement Algorithms with Implicit Differentiation
• SQ Lower Bounds for Learning Single Neurons with Massart Noise
• Meta-Reward-Net: Implicitly Differentiable Reward Learning for Preference-based Reinforcement Learning
• Average Sensitivity of Euclidean k-Clustering
• A theory of weight distribution-constrained learning
• Data augmentation for efficient learning from parametric experts
• Active Bayesian Causal Inference
• Template based Graph Neural Network with Optimal Transport Distances
• Outlier-Robust Sparse Estimation via Non-Convex Optimization
• Toward Understanding Privileged Features Distillation in Learning-to-Rank
• The Neural Covariance SDE: Shaped Infinite Depth-and-Width Networks at Initialization
• FP8 Quantization: The Power of the Exponent
• Maximizing Revenue under Market Shrinkage and Market Uncertainty
• UnfoldML: Cost-Aware and Uncertainty-Based Dynamic 2D Prediction for Multi-Stage Classification
• Structural Analysis of Branch-and-Cut and the Learnability of Gomory Mixed Integer Cuts
• DOPE: Doubly Optimistic and Pessimistic Exploration for Safe Reinforcement Learning
• Structure-Aware Image Segmentation with Homotopy Warping
• Deep Learning Methods for Proximal Inference via Maximum Moment Restriction
• On global convergence of ResNets: From finite to infinite width using linear parameterization
• Residual Multiplicative Filter Networks for Multiscale Reconstruction
• Reinforcement Learning with Non-Exponential Discounting
• Towards Trustworthy Automatic Diagnosis Systems by Emulating Doctors' Reasoning with Deep Reinforcement Learning
• On the symmetries of the synchronization problem in Cryo-EM: Multi-Frequency Vector Diffusion Maps on the Projective Plane
• A Theoretical View on Sparsely Activated Networks
• Pre-Train Your Loss: Easy Bayesian Transfer Learning with Informative Priors
• Implications of Model Indeterminacy for Explanations of Automated Decisions
• NOMAD: Nonlinear Manifold Decoders for Operator Learning
• Characterizing the Ventral Visual Stream with Response-Optimized Neural Encoding Models
• How Sampling Impacts the Robustness of Stochastic Neural Networks
• Forward-Backward Latent State Inference for Hidden Continuous-Time semi-Markov Chains
• Shape And Structure Preserving Differential Privacy
• On the Effectiveness of Lipschitz-Driven Rehearsal in Continual Learning
• Dynamic Pricing with Monotonicity Constraint under Unknown Parametric Demand Model
• Cross-Linked Unified Embedding for cross-modality representation learning
• Active Ranking without Strong Stochastic Transitivity
• ProtoVAE: A Trustworthy Self-Explainable Prototypical Variational Model
• The Mechanism of Prediction Head in Non-contrastive Self-supervised Learning
• Task Discovery: Finding the Tasks that Neural Networks Generalize on
• Chaotic Regularization and Heavy-Tailed Limits for Deterministic Gradient Descent
• LOT: Layer-wise Orthogonal Training on Improving l2 Certified Robustness
• Few-Shot Fast-Adaptive Anomaly Detection
• Learning dynamics of deep linear networks with multiple pathways
• Turbocharging Solution Concepts: Solving NEs, CEs and CCEs with Neural Equilibrium Solvers
• Multi-fidelity Monte Carlo: a pseudo-marginal approach
• Learning sparse features can lead to overfitting in neural networks
• Pushing the limits of fairness impossibility: Who's the fairest of them all?
• Neural Set Function Extensions: Learning with Discrete Functions in High Dimensions
• Zonotope Domains for Lagrangian Neural Network Verification
• Safety Guarantees for Neural Network Dynamic Systems via Stochastic Barrier Functions
• Online Bipartite Matching with Advice: Tight Robustness-Consistency Tradeoffs for the Two-Stage Model
• Improving Multi-Task Generalization via Regularizing Spurious Correlation
• Operative dimensions in unconstrained connectivity of recurrent neural networks
• Neural Differential Equations for Learning to Program Neural Nets Through Continuous Learning Rules
• Collaborative Linear Bandits with Adversarial Agents: Near-Optimal Regret Bounds
• Generating Training Data with Language Models: Towards Zero-Shot Language Understanding
• Differentially Private Graph Learning via Sensitivity-Bounded Personalized PageRank
• Towards Practical Few-shot Query Sets: Transductive Minimum Description Length Inference
• Randomized Channel Shuffling: Minimal-Overhead Backdoor Attack Detection without Clean Datasets
• MAgNet: Mesh Agnostic Neural PDE Solver
• Online Learning and Pricing for Network Revenue Management with Reusable Resources
• Learning Modular Simulations for Homogeneous Systems
• Instability and Local Minima in GAN Training with Kernel Discriminators
• On Computing Probabilistic Explanations for Decision Trees
• Distributed Optimization for Overparameterized Problems: Achieving Optimal Dimension Independent Communication Complexity
• Lost in Latent Space: Examining failures of disentangled models at combinatorial generalisation
• When Combinatorial Thompson Sampling meets Approximation Regret
• Test-Time Prompt Tuning for Zero-Shot Generalization in Vision-Language Models
• Detecting Abrupt Changes in Sequential Pairwise Comparison Data
• Sparse Fourier Backpropagation in Cryo-EM Reconstruction
• When Does Differentially Private Learning Not Suffer in High Dimensions?
• A Fast Scale-Invariant Algorithm for Non-negative Least Squares with Non-negative Data
• (Optimal) Online Bipartite Matching with Degree Information
• Learning from a Sample in Online Algorithms
• Data-Driven Conditional Robust Optimization
• Linear Label Ranking with Bounded Noise
• Estimation of Entropy in Constant Space with Improved Sample Complexity
• Escaping from the Barren Plateau via Gaussian Initializations in Deep Variational Quantum Circuits
• Expected Frequency Matrices of Elections: Computation, Geometry, and Preference Learning
• Robust Neural Posterior Estimation and Statistical Model Criticism
• CryptoGCN: Fast and Scalable Homomorphically Encrypted Graph Convolutional Network Inference
• The Missing Invariance Principle found -- the Reciprocal Twin of Invariant Risk Minimization
• MABSplit: Faster Forest Training Using Multi-Armed Bandits
• Marksman Backdoor: Backdoor Attacks with Arbitrary Target Class
• Collaborative Learning of Discrete Distributions under Heterogeneity and Communication Constraints
• Sample-Efficient Reinforcement Learning of Partially Observable Markov Games
• Phase transitions in when feedback is useful
• The Role of Baselines in Policy Gradient Optimization
• Autoformalization with Large Language Models
• Differentially Private Generalized Linear Models Revisited
• Learning to Follow Instructions in Text-Based Games
• On Learning and Refutation in Noninteractive Local Differential Privacy
• Cryptographic Hardness of Learning Halfspaces with Massart Noise
• Instance-optimal PAC Algorithms for Contextual Bandits
• Do Current Multi-Task Optimization Methods in Deep Learning Even Help?
• Prompt Certified Machine Unlearning with Randomized Gradient Smoothing and Quantization
• Unsupervised Reinforcement Learning with Contrastive Intrinsic Control
• Exact learning dynamics of deep linear networks with prior knowledge
• Self-Supervised Contrastive Pre-Training For Time Series via Time-Frequency Consistency
• Regret Bounds for Multilabel Classification in Sparse Label Regimes
• Characterizing Datapoints via Second-Split Forgetting
• Training language models to follow instructions with human feedback
• S4ND: Modeling Images and Videos as Multidimensional Signals with State Spaces
• Defining and Characterizing Reward Gaming
• Adversarial training for high-stakes reliability
• Semantic Probabilistic Layers for Neuro-Symbolic Learning
• WebShop: Towards Scalable Real-World Web Interaction with Grounded Language Agents
• Maximizing and Satisficing in Multi-armed Bandits with Graph Information
• Learning Probabilistic Models from Generator Latent Spaces with Hat EBM
• Spherical Channels for Modeling Atomic Interactions
• HyperTree Proof Search for Neural Theorem Proving
• Exploring the Latent Space of Autoencoders with Interventional Assays
• Root Cause Analysis of Failures in Microservices through Causal Discovery
• Graphein - a Python Library for Geometric Deep Learning and Network Analysis on Biomolecular Structures and Interaction Networks
• Finite-Sample Maximum Likelihood Estimation of Location
• BayesPCN: A Continually Learnable Predictive Coding Associative Memory
• On the detrimental effect of invariances in the likelihood for variational inference
• Learning to Compare Nodes in Branch and Bound with Graph Neural Networks
• Parameter-free Regret in High Probability with Heavy Tails
• Multi-Game Decision Transformers
• Structural Pruning via Latency-Saliency Knapsack
• The Query Complexity of Cake Cutting
• Best of Both Worlds Model Selection
• Pitfalls of Epistemic Uncertainty Quantification through Loss Minimisation
• New Definitions and Evaluations for Saliency Methods: Staying Intrinsic, Complete and Sound
• Memory safe computations with XLA compiler
• Learning NP-Hard Multi-Agent Assignment Planning using GNN: Inference on a Random Graph and Provable Auction-Fitted Q-learning
• Fairness in Federated Learning via Core-Stability
• Accelerating Certified Robustness Training via Knowledge Transfer
• Certifying Some Distributional Fairness with Subpopulation Decomposition
• A Few Expert Queries Suffices for Sample-Efficient RL with Resets and Linear Value Approximation
• Sublinear Algorithms for Hierarchical Clustering
• A Deep Reinforcement Learning Framework for Column Generation
• Logical Activation Functions: Logit-space equivalents of Probabilistic Boolean Operators
• EAGER: Asking and Answering Questions for Automatic Reward Shaping in Language-guided RL
• End-to-end Stochastic Optimization with Energy-based Model
• ReCo: Retrieve and Co-segment for Zero-shot Transfer
• Human-Robotic Prosthesis as Collaborating Agents for Symmetrical Walking
• Adaptive Interest for Emphatic Reinforcement Learning
• Chaotic Dynamics are Intrinsic to Neural Network Training with SGD
• Local Bayesian optimization via maximizing probability of descent
• Learning the Structure of Large Networked Systems Obeying Conservation Laws
• Near-Optimal No-Regret Learning Dynamics for General Convex Games
• The Impact of Task Underspecification in Evaluating Deep Reinforcement Learning
• A Practical, Progressively-Expressive GNN
• ELIGN: Expectation Alignment as a Multi-Agent Intrinsic Reward
• Provably tuning the ElasticNet across instances
• Fast Neural Kernel Embeddings for General Activations
• Evaluating Latent Space Robustness and Uncertainty of EEG-ML Models under Realistic Distribution Shifts
• Simple and Optimal Greedy Online Contention Resolution Schemes
• Modeling Transitivity and Cyclicity in Directed Graphs via Binary Code Box Embeddings
• Planning to the Information Horizon of BAMDPs via Epistemic State Abstraction
• Decoupled Context Processing for Context Augmented Language Modeling
• Efficiency Ordering of Stochastic Gradient Descent
• Robust Streaming PCA
• Learning Partial Equivariances From Data
• [Re] Lifting 2D StyleGAN for 3D-Aware Face Generation
• FiLM: Frequency improved Legendre Memory Model for Long-term Time Series Forecasting
• Unsupervised Causal Generative Understanding of Images
• Curriculum Reinforcement Learning using Optimal Transport via Gradual Domain Adaptation
• Fair Ranking with Noisy Protected Attributes
• Independence Testing-Based Approach to Causal Discovery under Measurement Error and Linear Non-Gaussian Models
• Zero-Sum Stochastic Stackelberg Games
• Can Hybrid Geometric Scattering Networks Help Solve the Maximum Clique Problem?
• NeurOLight: A Physics-Agnostic Neural Operator Enabling Parametric Photonic Device Simulation
• Mining Multi-Label Samples from Single Positive Labels
• Why So Pessimistic? Estimating Uncertainties for Offline RL through Ensembles, and Why Their Independence Matters
• Efficient Phi-Regret Minimization in Extensive-Form Games via Online Mirror Descent
• Exponential Family Model-Based Reinforcement Learning via Score Matching
• Object Scene Representation Transformer
• Geometric Order Learning for Rank Estimation
• Learning with convolution and pooling operations in kernel methods
• Dataset Distillation using Neural Feature Regression
• Influencing Long-Term Behavior in Multiagent Reinforcement Learning
• Model-based Safe Deep Reinforcement Learning via a Constrained Proximal Policy Optimization Algorithm
• Unifying and Boosting Gradient-Based Training-Free Neural Architecture Search
• Learning Contrastive Embedding in Low-Dimensional Space
• Exploring Example Influence in Continual Learning
• JAWS: Auditing Predictive Uncertainty Under Covariate Shift
• One for All: Simultaneous Metric and Preference Learning over Multiple Users
• Paraphrasing Is All You Need for Novel Object Captioning
• Augmentations in Hypergraph Contrastive Learning: Fabricated and Generative
• Multiview Human Body Reconstruction from Uncalibrated Cameras
• FairVFL: A Fair Vertical Federated Learning Framework with Contrastive Adversarial Learning
• Empirical Gateaux Derivatives for Causal Inference
• AgraSSt: Approximate Graph Stein Statistics for Interpretable Assessment of Implicit Graph Generators
• Benefits of Permutation-Equivariance in Auction Mechanisms
• Learning Active Camera for Multi-Object Navigation
• Toward Efficient Robust Training against Union of $\ell_p$ Threat Models
• Mask Matching Transformer for Few-Shot Segmentation
• A Unified Hard-Constraint Framework for Solving Geometrically Complex PDEs
• Symplectic Spectrum Gaussian Processes: Learning Hamiltonians from Noisy and Sparse Data
• GREED: A Neural Framework for Learning Graph Distance Functions
• Understanding Cross-Domain Few-Shot Learning Based on Domain Similarity and Few-Shot Difficulty
• Consistent Sufficient Explanations and Minimal Local Rules for explaining the decision of any classifier or regressor
• DaDA: Distortion-aware Domain Adaptation for Unsupervised Semantic Segmentation
• Learning Optical Flow from Continuous Spike Streams
• Retrospective Adversarial Replay for Continual Learning
• Adversarial Auto-Augment with Label Preservation: A Representation Learning Principle Guided Approach
• On Feature Learning in the Presence of Spurious Correlations
• Explaining Preferences with Shapley Values
• Privacy Induces Robustness: Information-Computation Gaps and Sparse Mean Estimation
• ReFactor GNNs: Revisiting Factorisation-based Models from a Message-Passing Perspective
• Block-Recurrent Transformers
• Hamiltonian Latent Operators for content and motion disentanglement in image sequences
• Learning (Very) Simple Generative Models Is Hard
• Understanding the Generalization Benefit of Normalization Layers: Sharpness Reduction
• A Non-asymptotic Analysis of Non-parametric Temporal-Difference Learning
• Rethinking Generalization in Few-Shot Classification
• VectorAdam for Rotation Equivariant Geometry Optimization
• Keypoint-Guided Optimal Transport with Applications in Heterogeneous Domain Adaptation
• Supervising the Multi-Fidelity Race of Hyperparameter Configurations
• Trajectory of Mini-Batch Momentum: Batch Size Saturation and Convergence in High Dimensions
• Single Model Uncertainty Estimation via Stochastic Data Centering
• CogView2: Faster and Better Text-to-Image Generation via Hierarchical Transformers
• Graph Neural Networks with Adaptive Readouts
• Adaptive Distribution Calibration for Few-Shot Learning with Hierarchical Optimal Transport
• Towards Reliable Simulation-Based Inference with Balanced Neural Ratio Estimation
• Beyond Mahalanobis Distance for Textual OOD Detection
• Tensor Program Optimization with Probabilistic Programs
• VICE: Variational Interpretable Concept Embeddings
• Learning single-index models with shallow neural networks
• Near-Optimal Randomized Exploration for Tabular Markov Decision Processes
• Understanding Non-linearity in Graph Neural Networks from the Bayesian-Inference Perspective
• LOG: Active Model Adaptation for Label-Efficient OOD Generalization
• Structural Knowledge Distillation for Object Detection
• Semantic uncertainty intervals for disentangled latent spaces
• Uni[MASK]: Unified Inference in Sequential Decision Problems
• Bayesian Optimistic Optimization: Optimistic Exploration for Model-based Reinforcement Learning
• Invertible Monotone Operators for Normalizing Flows
• A Transformer-Based Object Detector with Coarse-Fine Crossing Representations
• Distinguishing Learning Rules with Brain Machine Interfaces
• Asynchronous Actor-Critic for Multi-Agent Reinforcement Learning
• Expected Improvement for Contextual Bandits
• BEVFusion: A Simple and Robust LiDAR-Camera Fusion Framework
• Graph Neural Network Bandits
• Mean Estimation with User-level Privacy under Data Heterogeneity
• Precise Regret Bounds for Log-loss via a Truncated Bayesian Algorithm
• ViewFool: Evaluating the Robustness of Visual Recognition to Adversarial Viewpoints
• LobsDICE: Offline Learning from Observation via Stationary Distribution Correction Estimation
• 360-MLC: Multi-view Layout Consistency for Self-training and Hyper-parameter Tuning
• Reduced Representation of Deformation Fields for Effective Non-rigid Shape Matching
• Explain My Surprise: Learning Efficient Long-Term Memory by predicting uncertain outcomes
• A Simple and Optimal Policy Design for Online Learning with Safety against Heavy-tailed Risk
• Policy Optimization with Linear Temporal Logic Constraints
• Scaling & Shifting Your Features: A New Baseline for Efficient Model Tuning
• Decomposed Knowledge Distillation for Class-Incremental Semantic Segmentation
• On the Theoretical Properties of Noise Correlation in Stochastic Optimization
• NCP: Neural Correspondence Prior for Effective Unsupervised Shape Matching
• ComENet: Towards Complete and Efficient Message Passing for 3D Molecular Graphs
• On Divergence Measures for Bayesian Pseudocoresets
• Alleviating the Sample Selection Bias in Few-shot Learning by Removing Projection to the Centroid
• Can Push-forward Generative Models Fit Multimodal Distributions?
• Posterior and Computational Uncertainty in Gaussian Processes
• MORA: Improving Ensemble Robustness Evaluation with Model Reweighing Attack
• Learning to Sample and Aggregate: Few-shot Reasoning over Temporal Knowledge Graphs
• Advancing Model Pruning via Bi-level Optimization
• An Algorithm for Learning Switched Linear Dynamics from Data
• Batch Bayesian optimisation via density-ratio estimation with guarantees
• An Adaptive Kernel Approach to Federated Learning of Heterogeneous Causal Effects
• Structuring Uncertainty for Fine-Grained Sampling in Stochastic Segmentation Networks
• Multi-objective Deep Data Generation with Correlated Property Control
• Expectation-Maximization Contrastive Learning for Compact Video-and-Language Representations
• Learning Debiased Classifier with Biased Committee
• Surprising Instabilities in Training Deep Networks and a Theoretical Analysis
• Capturing Failures of Large Language Models via Human Cognitive Biases
• Nonnegative Tensor Completion via Integer Optimization
• Equivariant Networks for Crystal Structures
• LieGG: Studying Learned Lie Group Generators
• On-Demand Sampling: Learning Optimally from Multiple Distributions
• A Communication-efficient Algorithm with Linear Convergence for Federated Minimax Learning
• Robust Model Selection and Nearly-Proper Learning for GMMs
• Combining Explicit and Implicit Regularization for Efficient Learning in Deep Networks
• A Unified Framework for Alternating Offline Model Training and Policy Learning
• Automatic Differentiation of Programs with Discrete Randomness
• Thinned random measures for sparse graphs with overlapping communities
• Hybrid Neural Autoencoders for Stimulus Encoding in Visual and Other Sensory Neuroprostheses
• Tight Analysis of Extra-gradient and Optimistic Gradient Methods For Nonconvex Minimax Problems
• Amortized Inference for Causal Structure Learning
• Staircase Attention for Recurrent Processing of Sequences
• A Multi-Resolution Framework for U-Nets with Applications to Hierarchical VAEs
• Concept Embedding Models: Beyond the Accuracy-Explainability Trade-Off
• Biologically plausible solutions for spiking networks with efficient coding
• Algorithms with Prediction Portfolios
• SAGDA: Achieving $\mathcal{O}(\epsilon^{-2})$ Communication Complexity in Federated Min-Max Learning
• Deep Compression of Pre-trained Transformer Models
• Beyond neural scaling laws: beating power law scaling via data pruning
• Subgroup Robustness Grows On Trees: An Empirical Baseline Investigation
• Learning Optimal Flows for Non-Equilibrium Importance Sampling
• Incentivizing Combinatorial Bandit Exploration
• A Simple Decentralized Cross-Entropy Method
• Neural Abstractions
• Learning Dense Object Descriptors from Multiple Views for Low-shot Category Generalization
• Flowification: Everything is a normalizing flow
• Non-monotonic Resource Utilization in the Bandits with Knapsacks Problem
• Evaluating Robustness to Dataset Shift via Parametric Robustness Sets
• Generative Time Series Forecasting with Diffusion, Denoise, and Disentanglement
• Private and Communication-Efficient Algorithms for Entropy Estimation
• Kernel Multimodal Continuous Attention
• Stars: Tera-Scale Graph Building for Clustering and Learning
• Anonymous Bandits for Multi-User Systems
• Understanding Deep Contrastive Learning via Coordinate-wise Optimization
• Stochastic Halpern Iteration with Variance Reduction for Stochastic Monotone Inclusions
• PALMER: Perception - Action Loop with Memory for Long-Horizon Planning
• Curious Exploration via Structured World Models Yields Zero-Shot Object Manipulation
• Finite Sample Analysis Of Dynamic Regression Parameter Learning
• Adaptive Stochastic Variance Reduction for Non-convex Finite-Sum Minimization
• Improved Coresets for Euclidean $k$-Means
• Data-Efficient Pipeline for Offline Reinforcement Learning with Limited Data
• Distributed Influence-Augmented Local Simulators for Parallel MARL in Large Networked Systems
• Neural Temporal Walks: Motif-Aware Representation Learning on Continuous-Time Dynamic Graphs
• Grounded Video Situation Recognition
• Learning to Scaffold: Optimizing Model Explanations for Teaching
• Public Wisdom Matters! Discourse-Aware Hyperbolic Fourier Co-Attention for Social Text Classification
• Efficient Methods for Non-stationary Online Learning
• Sustainable Online Reinforcement Learning for Auto-bidding
• Effectiveness of Vision Transformer for Fast and Accurate Single-Stage Pedestrian Detection
• On Analyzing Generative and Denoising Capabilities of Diffusion-based Deep Generative Models
• Isometric 3D Adversarial Examples in the Physical World
• The Hessian Screening Rule
• Measuring Data Reconstruction Defenses in Collaborative Inference Systems
• A Stochastic Linearized Augmented Lagrangian Method for Decentralized Bilevel Optimization
• Kernel Memory Networks: A Unifying Framework for Memory Modeling
• A Neural Pre-Conditioning Active Learning Algorithm to Reduce Label Complexity
• Flexible Diffusion Modeling of Long Videos
• Learning Structure from the Ground up---Hierarchical Representation Learning by Chunking
• Meta-Complementing the Semantics of Short Texts in Neural Topic Models
• Robust Feature-Level Adversaries are Interpretability Tools
• Knowledge-Aware Bayesian Deep Topic Model
• GStarX: Explaining Graph Neural Networks with Structure-Aware Cooperative Games
• Quantum Algorithms for Sampling Log-Concave Distributions and Estimating Normalizing Constants
• Nearly Optimal Best-of-Both-Worlds Algorithms for Online Learning with Feedback Graphs
• FourierNets enable the design of highly non-local optical encoders for computational imaging
• TVLT: Textless Vision-Language Transformer
• No Free Lunch from Deep Learning in Neuroscience: A Case Study through Models of the Entorhinal-Hippocampal Circuit
• Retaining Knowledge for Learning with Dynamic Definition
• XTC: Extreme Compression for Pre-trained Transformers Made Simple and Efficient
• PAC: Assisted Value Factorization with Counterfactual Predictions in Multi-Agent Reinforcement Learning
• Meta-DMoE: Adapting to Domain Shift by Meta-Distillation from Mixture-of-Experts
• Compositional Generalization in Unsupervised Compositional Representation Learning: A Study on Disentanglement and Emergent Language
• Fairness without Demographics through Knowledge Distillation
• Deep Bidirectional Language-Knowledge Graph Pretraining
• Rethinking Value Function Learning for Generalization in Reinforcement Learning
• Instance-Dependent Near-Optimal Policy Identification in Linear MDPs via Online Experiment Design
• Exposing and Exploiting Fine-Grained Block Structures for Fast and Accurate Sparse Training
• Parameters or Privacy: A Provable Tradeoff Between Overparameterization and Membership Inference
• Efficient Dataset Distillation using Random Feature Approximation
• Locally Hierarchical Auto-Regressive Modeling for Image Generation
• Interaction-Grounded Learning with Action-Inclusive Feedback
• AdaFocal: Calibration-aware Adaptive Focal Loss
• Convergence for score-based generative modeling with polynomial complexity
• Toward Robust Spiking Neural Network Against Adversarial Perturbation
• Layer Freezing & Data Sieving: Missing Pieces of a Generic Framework for Sparse Training
• Efficiently Computing Local Lipschitz Constants of Neural Networks via Bound Propagation
• $\alpha$-ReQ : Assessing Representation Quality in Self-Supervised Learning by measuring eigenspectrum decay
• Bounded-Regret MPC via Perturbation Analysis: Prediction Error, Constraints, and Nonlinearity
• NaturalProver: Grounded Mathematical Proof Generation with Language Models
• Predictive Querying for Autoregressive Neural Sequence Models
• Differentially Private Linear Sketches: Efficient Implementations and Applications
• Probable Domain Generalization via Quantile Risk Minimization
• Embed and Emulate: Learning to estimate parameters of dynamical systems with uncertainty quantification
• Minimax Optimal Online Imitation Learning via Replay Estimation
• Subspace Recovery from Heterogeneous Data with Non-isotropic Noise
• Transferring Fairness under Distribution Shifts via Fair Consistency Regularization
• Exploring the Whole Rashomon Set of Sparse Decision Trees
• On Image Segmentation With Noisy Labels: Characterization and Volume Properties of the Optimal Solutions to Accuracy and Dice
• AutoML Two-Sample Test
• Efficient Scheduling of Data Augmentation for Deep Reinforcement Learning
• Which Explanation Should I Choose? A Function Approximation Perspective to Characterizing Post Hoc Explanations
• Sampling from Log-Concave Distributions with Infinity-Distance Guarantees
• Decentralized Gossip-Based Stochastic Bilevel Optimization over Communication Networks
• Distributional Reinforcement Learning for Risk-Sensitive Policies
• Causal Discovery in Heterogeneous Environments Under the Sparse Mechanism Shift Hypothesis
• Promising or Elusive? Unsupervised Object Segmentation from Real-world Single Images
• Unsupervised Domain Adaptation for Semantic Segmentation using Depth Distribution
• Data-Efficient Structured Pruning via Submodular Optimization
• Structured Energy Network As a Loss
• Iso-Dream: Isolating and Leveraging Noncontrollable Visual Dynamics in World Models
• Improving Barely Supervised Learning by Discriminating Unlabeled Samples with Super-Class
• Rethinking and Improving Robustness of Convolutional Neural Networks: a Shapley Value-based Approach in Frequency Domain
• Exploring evolution-aware & -free protein language models as protein function predictors
• Boosting the Performance of Generic Deep Neural Network Frameworks with Log-supermodular CRFs
• On the Tradeoff Between Robustness and Fairness
• Learning to Reason with Neural Networks: Generalization, Unseen Data and Boolean Measures
• Causality-driven Hierarchical Structure Discovery for Reinforcement Learning
• Are AlphaZero-like Agents Robust to Adversarial Perturbations?
• Thinking Outside the Ball: Optimal Learning with Gradient Descent for Generalized Linear Stochastic Convex Optimization
• Pluralistic Image Completion with Gaussian Mixture Models
• Generalization Analysis on Learning with a Concurrent Verifier
• Receding Horizon Inverse Reinforcement Learning
• Learning to Share in Networked Multi-Agent Reinforcement Learning
• FIRE: Semantic Field of Words Represented as Non-Linear Functions
• Perceptual Attacks of No-Reference Image Quality Models with Human-in-the-Loop
• ProtoX: Explaining a Reinforcement Learning Agent via Prototyping
• Pyramid Attention For Source Code Summarization
• Taming Fat-Tailed (“Heavier-Tailed” with Potentially Infinite Variance) Noise in Federated Learning
• Maximum a posteriori natural scene reconstruction from retinal ganglion cells with deep denoiser priors
• DigGAN: Discriminator gradIent Gap Regularization for GAN Training with Limited Data
• DNA: Proximal Policy Optimization with a Dual Network Architecture
• Will Bilevel Optimizers Benefit from Loops
• Micro and Macro Level Graph Modeling for Graph Variational Auto-Encoders
• Redeeming intrinsic rewards via constrained optimization
• Target alignment in truncated kernel ridge regression
• Queue Up Your Regrets: Achieving the Dynamic Capacity Region of Multiplayer Bandits
• Mismatched No More: Joint Model-Policy Optimization for Model-Based RL
• Dynamic Sparse Network for Time Series Classification: Learning What to “See”
• Stochastic Second-Order Methods Improve Best-Known Sample Complexity of SGD for Gradient-Dominated Functions
• Constrained Predictive Coding as a Biologically Plausible Model of the Cortical Hierarchy
• Perturbation Learning Based Anomaly Detection
• Hierarchical Graph Transformer with Adaptive Node Sampling
• LogiGAN: Learning Logical Reasoning via Adversarial Pre-training
• Learning Causally Invariant Representations for Out-of-Distribution Generalization on Graphs
• Structure-Preserving 3D Garment Modeling with Neural Sewing Machines
• Improved Bounds on Neural Complexity for Representing Piecewise Linear Functions
• On the Limitations of Stochastic Pre-processing Defenses
• ResQ: A Residual Q Function-based Approach for Multi-Agent Reinforcement Learning Value Factorization
• Global Convergence and Stability of Stochastic Gradient Descent
• Delving into Out-of-Distribution Detection with Vision-Language Representations
• Recruitment Strategies That Take a Chance
• Inference and Sampling for Archimax Copulas
• Text Classification with Born's Rule
• Cluster and Aggregate: Face Recognition with Large Probe Set
• VTC-LFC: Vision Transformer Compression with Low-Frequency Components
• Lipschitz Bandits with Batched Feedback
• Formulating Robustness Against Unforeseen Attacks
• Randomized Message-Interception Smoothing: Gray-box Certificates for Graph Neural Networks
• Subspace clustering in high-dimensions: Phase transitions & Statistical-to-Computational gap
• CageNeRF: Cage-based Neural Radiance Field for Generalized 3D Deformation and Animation
• Sparse2Dense: Learning to Densify 3D Features for 3D Object Detection
• Non-Gaussian Tensor Programs
• Understanding the Eluder Dimension
• Semi-supervised Semantic Segmentation with Prototype-based Consistency Regularization
• Local Linear Convergence of Gradient Methods for Subspace Optimization via Strict Complementarity
• Simulation-guided Beam Search for Neural Combinatorial Optimization
• Quo Vadis: Is Trajectory Forecasting the Key Towards Long-Term Multi-Object Tracking?
• Meta-Reinforcement Learning with Self-Modifying Networks
• Respecting Transfer Gap in Knowledge Distillation
• What is Where by Looking: Weakly-Supervised Open-World Phrase-Grounding without Text Inputs
• TarGF: Learning Target Gradient Field to Rearrange Objects without Explicit Goal Specification
• One-Inlier is First: Towards Efficient Position Encoding for Point Cloud Registration
• I2DFormer: Learning Image to Document Attention for Zero-Shot Image Classification
• Sharing Knowledge for Meta-learning with Feature Descriptions
• Large-batch Optimization for Dense Visual Predictions: Training Faster R-CNN in 4.2 Minutes
• Continual Learning with Evolving Class Ontologies
• Quasi-Newton Methods for Saddle Point Problems
• TANGO: Text-driven Photorealistic and Robust 3D Stylization via Lighting Decomposition
• Asymptotic Properties for Bayesian Neural Network in Besov Space
• Planning for Sample Efficient Imitation Learning
• Peripheral Vision Transformer
• Multi-block-Single-probe Variance Reduced Estimator for Coupled Compositional Optimization
• HSDF: Hybrid Sign and Distance Field for Modeling Surfaces with Arbitrary Topologies
• Approximate Secular Equations for the Cubic Regularization Subproblem
• Faster Stochastic Algorithms for Minimax Optimization under Polyak-{\L}ojasiewicz Condition
• Unsupervised Learning of Equivariant Structure from Sequences
• Inception Transformer
• Signal Recovery with Non-Expansive Generative Network Priors
• Counterfactual harm
• Posterior Collapse of a Linear Latent Variable Model
• Harmonizing the object recognition strategies of deep neural networks with humans
• When to Make Exceptions: Exploring Language Models as Accounts of Human Moral Judgment
• Exploit Reward Shifting in Value-Based Deep-RL: Optimistic Curiosity-Based Exploration and Conservative Exploitation via Linear Reward Shaping
• Model-Based Imitation Learning for Urban Driving
• OnePose++: Keypoint-Free One-Shot Object Pose Estimation without CAD Models
• ELIAS: End-to-End Learning to Index and Search in Large Output Spaces
• QUARK: Controllable Text Generation with Reinforced Unlearning
• Self-Supervised Aggregation of Diverse Experts for Test-Agnostic Long-Tailed Recognition
• Finding Differences Between Transformers and ConvNets Using Counterfactual Simulation Testing
• Anticipating Performativity by Predicting from Predictions
• Fast Vision Transformers with HiLo Attention
• OpenAUC: Towards AUC-Oriented Open-Set Recognition
• Exploring the Algorithm-Dependent Generalization of AUPRC Optimization with List Stability
• Efficient Spatially Sparse Inference for Conditional GANs and Diffusion Models
• Differentiable Analog Quantum Computing for Optimization and Control
• Losses Can Be Blessings: Routing Self-Supervised Speech Representations Towards Efficient Multilingual and Multitask Speech Processing
• Monte Carlo Tree Descent for Black-Box Optimization
• On Reinforcement Learning and Distribution Matching for Fine-Tuning Language Models with no Catastrophic Forgetting
• Robust Imitation of a Few Demonstrations with a Backwards Model
• AMP: Automatically Finding Model Parallel Strategies with Heterogeneity Awareness
• Communication Acceleration of Local Gradient Methods via an Accelerated Primal-Dual Algorithm with an Inexact Prox
• Performative Power
• SatMAE: Pre-training Transformers for Temporal and Multi-Spectral Satellite Imagery
• Benign, Tempered, or Catastrophic: Toward a Refined Taxonomy of Overfitting
• The Power and Limitation of Pretraining-Finetuning for Linear Regression under Covariate Shift
• Confident Approximate Policy Iteration for Efficient Local Planning in $q^\pi$-realizable MDPs
• Unpacking Reward Shaping: Understanding the Benefits of Reward Engineering on Sample Complexity
• Society of Agents: Regret Bounds of Concurrent Thompson Sampling
• Exploring Length Generalization in Large Language Models
• Unsupervised Learning for Combinatorial Optimization with Principled Objective Relaxation
• GPT3.int8(): 8-bit Matrix Multiplication for Transformers at Scale
• Lottery Tickets on a Data Diet: Finding Initializations with Sparse Trainable Networks
• Revisiting Sparse Convolutional Model for Visual Recognition
• Temporal Latent Bottleneck: Synthesis of Fast and Slow Processing Mechanisms in Sequence Learning
• MoCoDA: Model-based Counterfactual Data Augmentation
• Beyond Adult and COMPAS: Fair Multi-Class Prediction via Information Projection
• On the generalization of learning algorithms that do not converge
• Capturing Graphs with Hypo-Elliptic Diffusions
• Hypothesis Testing for Differentially Private Linear Regression
• Recurrent Convolutional Neural Networks Learn Succinct Learning Algorithms
• AutoST: Towards the Universal Modeling of Spatio-temporal Sequences
• SoLar: Sinkhorn Label Refinery for Imbalanced Partial-Label Learning
• ESCADA: Efficient Safety and Context Aware Dose Allocation for Precision Medicine
• Explicit Tradeoffs between Adversarial and Natural Distributional Robustness
• Generalization Bounds for Gradient Methods via Discrete and Continuous Prior
• CascadeXML: Rethinking Transformers for End-to-end Multi-resolution Training in Extreme Multi-label Classification
• BYOL-Explore: Exploration by Bootstrapped Prediction
• Ordered Subgraph Aggregation Networks
• Where do Models go Wrong? Parameter-Space Saliency Maps for Explainability
• Label Noise in Adversarial Training: A Novel Perspective to Study Robust Overfitting
• Risk Bounds of Multi-Pass SGD for Least Squares in the Interpolation Regime
• Text-Adaptive Multiple Visual Prototype Matching for Video-Text Retrieval
• An In-depth Study of Stochastic Backpropagation
• Tractable Optimality in Episodic Latent MABs
• Meta-Query-Net: Resolving Purity-Informativeness Dilemma in Open-set Active Learning
• Improving Certified Robustness via Statistical Learning with Logical Reasoning
• Online Decision Mediation
• Deep Differentiable Logic Gate Networks
• Double Bubble, Toil and Trouble: Enhancing Certified Robustness through Transitivity
• Associating Objects and Their Effects in Video through Coordination Games
• Finite-Time Regret of Thompson Sampling Algorithms for Exponential Family Multi-Armed Bandits
• Precise Learning Curves and Higher-Order Scalings for Dot-product Kernel Regression
• Quantifying Statistical Significance of Neural Network-based Image Segmentation by Selective Inference
• Multi-block Min-max Bilevel Optimization with Applications in Multi-task Deep AUC Maximization
• Agreement-on-the-line: Predicting the Performance of Neural Networks under Distribution Shift
• Neural Conservation Laws: A Divergence-Free Perspective
• Sparse Hypergraph Community Detection Thresholds in Stochastic Block Model
• Understanding and Extending Subgraph GNNs by Rethinking Their Symmetries
• Latent Hierarchical Causal Structure Discovery with Rank Constraints
• Task-Agnostic Graph Explanations
• ZSON: Zero-Shot Object-Goal Navigation using Multimodal Goal Embeddings
• Towards Optimal Communication Complexity in Distributed Non-Convex Optimization
• Fast Bayesian Coresets via Subsampling and Quasi-Newton Refinement
• Optimal Rates for Regularized Conditional Mean Embedding Learning
• Are All Losses Created Equal: A Neural Collapse Perspective
• Scalable Representation Learning in Linear Contextual Bandits with Constant Regret Guarantees
• What You See is What You Get: Principled Deep Learning via Distributional Generalization
• Knowledge Distillation: Bad Models Can Be Good Role Models
• Fine-Tuning Pre-Trained Language Models Effectively by Optimizing Subnetworks Adaptively
• Lifelong Neural Predictive Coding: Learning Cumulatively Online without Forgetting
• Rare Gems: Finding Lottery Tickets at Initialization
• Hidden Progress in Deep Learning: SGD Learns Parities Near the Computational Limit
• Neural Approximation of Graph Topological Features
• Near-Optimal Regret Bounds for Multi-batch Reinforcement Learning
• Surprise Minimizing Multi-Agent Learning with Energy-based Models
• Sparse Structure Search for Delta Tuning
• Stability and Generalization for Markov Chain Stochastic Gradient Methods
• Leveraging Factored Action Spaces for Efficient Offline Reinforcement Learning in Healthcare
• Discovery of Single Independent Latent Variable
• MoGDE: Boosting Mobile Monocular 3D Object Detection with Ground Depth Estimation
• Compressible-composable NeRF via Rank-residual Decomposition
• Asymmetric Temperature Scaling Makes Larger Networks Teach Well Again
• DPM-Solver: A Fast ODE Solver for Diffusion Probabilistic Model Sampling in Around 10 Steps
• Uni-Perceiver-MoE: Learning Sparse Generalist Models with Conditional MoEs
• Neural Shape Deformation Priors
• Hierarchical Channel-spatial Encoding for Communication-efficient Collaborative Learning
• Debugging and Explaining Metric Learning Approaches: An Influence Function Based Perspective
• Factuality Enhanced Language Models for Open-Ended Text Generation
• Learn to Match with No Regret: Reinforcement Learning in Markov Matching Markets
• A Simple and Provably Efficient Algorithm for Asynchronous Federated Contextual Linear Bandits
• MaskTune: Mitigating Spurious Correlations by Forcing to Explore
• Scalable Sensitivity and Uncertainty Analyses for Causal-Effect Estimates of Continuous-Valued Interventions
• Additive MIL: Intrinsically Interpretable Multiple Instance Learning for Pathology
• Reconstructing Training Data From Trained Neural Networks
• Use-Case-Grounded Simulations for Explanation Evaluation
• Differentiable hierarchical and surrogate gradient search for spiking neural networks
• CalFAT: Calibrated Federated Adversarial Training with Label Skewness
• Cluster Randomized Designs for One-Sided Bipartite Experiments
• Multi-Sample Training for Neural Image Compression
• On the Parameterization and Initialization of Diagonal State Space Models
• Solving Quantitative Reasoning Problems with Language Models
• Learnable Polyphase Sampling for Shift Invariant and Equivariant Convolutional Networks
• D^2NeRF: Self-Supervised Decoupling of Dynamic and Static Objects from a Monocular Video
• Semi-Supervised Video Salient Object Detection Based on Uncertainty-Guided Pseudo Labels
• C2FAR: Coarse-to-Fine Autoregressive Networks for Precise Probabilistic Forecasting
• SizeShiftReg: a Regularization Method for Improving Size-Generalization in Graph Neural Networks
• Squeezeformer: An Efficient Transformer for Automatic Speech Recognition
• Adversarial Attack on Attackers: Post-Process to Mitigate Black-Box Score-Based Query Attacks
• Generalizing Bayesian Optimization with Decision-theoretic Entropies
• Dict-TTS: Learning to Pronounce with Prior Dictionary Knowledge for Text-to-Speech
• The Unreasonable Effectiveness of Fully-Connected Layers for Low-Data Regimes
• Unsupervised Object Detection Pretraining with Joint Object Priors Generation and Detector Learning
• Learning Chaotic Dynamics in Dissipative Systems
• MonoSDF: Exploring Monocular Geometric Cues for Neural Implicit Surface Reconstruction
• SeqPATE: Differentially Private Text Generation via Knowledge Distillation
• DENSE: Data-Free One-Shot Federated Learning
• Sym-NCO: Leveraging Symmetricity for Neural Combinatorial Optimization
• Is $L^2$ Physics Informed Loss Always Suitable for Training Physics Informed Neural Network?
• Hiding Images in Deep Probabilistic Models
• Factored Adaptation for Non-Stationary Reinforcement Learning
• Optimal Algorithms for Decentralized Stochastic Variational Inequalities
• Semi-supervised Vision Transformers at Scale
• Deep Model Reassembly
• Your Transformer May Not be as Powerful as You Expect
• InsPro: Propagating Instance Query and Proposal for Online Video Instance Segmentation
• Linear tree shap
• Delving into Sequential Patches for Deepfake Detection
• Untargeted Backdoor Watermark: Towards Harmless and Stealthy Dataset Copyright Protection
• ClimbQ: Class Imbalanced Quantization Enabling Robustness on Efficient Inferences
• Learning Latent Seasonal-Trend Representations for Time Series Forecasting
• Posterior Refinement Improves Sample Efficiency in Bayesian Neural Networks
• Back Razor: Memory-Efficient Transfer Learning by Self-Sparsified Backpropagation
• DreamShard: Generalizable Embedding Table Placement for Recommender Systems
• Dataset Distillation via Factorization
• Video Diffusion Models
• Theseus: A Library for Differentiable Nonlinear Optimization
• Decoupling Features in Hierarchical Propagation for Video Object Segmentation
• RankFeat: Rank-1 Feature Removal for Out-of-distribution Detection
• Explainable Reinforcement Learning via Model Transforms
• Matryoshka Representation Learning
• VoxGRAF: Fast 3D-Aware Image Synthesis with Sparse Voxel Grids
• Decoupling Classifier for Boosting Few-shot Object Detection and Instance Segmentation
• MetaMask: Revisiting Dimensional Confounder for Self-Supervised Learning
• LGDN: Language-Guided Denoising Network for Video-Language Modeling
• PyramidCLIP: Hierarchical Feature Alignment for Vision-language Model Pretraining
• Divide and Contrast: Source-free Domain Adaptation via Adaptive Contrastive Learning
• Adapting Self-Supervised Vision Transformers by Probing Attention-Conditioned Masking Consistency
• Flexible Neural Image Compression via Code Editing
• Learning Physics Constrained Dynamics Using Autoencoders
• Active Learning with Neural Networks: Insights from Nonparametric Statistics
• Understanding Robust Learning through the Lens of Representation Similarities
• Beyond Separability: Analyzing the Linear Transferability of Contrastive Representations to Related Subpopulations
• Few-shot Task-agnostic Neural Architecture Search for Distilling Large Language Models
• Zero-Shot Video Question Answering via Frozen Bidirectional Language Models
• Outsourcing Training without Uploading Data via Efficient Collaborative Open-Source Sampling
• Measuring and Reducing Model Update Regression in Structured Prediction for NLP
• Coordinates Are NOT Lonely - Codebook Prior Helps Implicit Neural 3D representations
• Multitasking Models are Robust to Structural Failure: A Neural Model for Bilingual Cognitive Reserve
• A Policy-Guided Imitation Approach for Offline Reinforcement Learning
• Asymptotics of smoothed Wasserstein distances in the small noise regime
• Finite-Time Last-Iterate Convergence for Learning in Multi-Player Games
• CARD: Classification and Regression Diffusion Models
• GraphDE: A Generative Framework for Debiased Learning and Out-of-Distribution Detection on Graphs
• Unlabelled Sample Compression Schemes for Intersection-Closed Classes and Extremal Classes
• Concentration of Data Encoding in Parameterized Quantum Circuits
• Learning Efficient Vision Transformers via Fine-Grained Manifold Distillation
• M$^4$I: Multi-modal Models Membership Inference
• Beyond accuracy: generalization properties of bio-plausible temporal credit assignment rules
• VLMo: Unified Vision-Language Pre-Training with Mixture-of-Modality-Experts
• Pre-Trained Language Models for Interactive Decision-Making
• Learning from Label Proportions by Learning with Label Noise
• A Closer Look at Offline RL Agents
• Beyond spectral gap: the role of the topology in decentralized learning
• A permutation-free kernel two-sample test
• C-Mixup: Improving Generalization in Regression
• Generalizing Consistent Multi-Class Classification with Rejection to be Compatible with Arbitrary Losses
• Efficient Multi-agent Communication via Self-supervised Information Aggregation
• EfficientFormer: Vision Transformers at MobileNet Speed
• Pseudo-Riemannian Graph Convolutional Networks
• Fast Algorithms for Packing Proportional Fairness and its Dual
• Optimistic Posterior Sampling for Reinforcement Learning with Few Samples and Tight Guarantees
• Training Scale-Invariant Neural Networks on the Sphere Can Happen in Three Regimes
• Reincarnating Reinforcement Learning: Reusing Prior Computation to Accelerate Progress
• Active Exploration for Inverse Reinforcement Learning
• UniGAN: Reducing Mode Collapse in GANs using a Uniform Generator
• Diffusion Curvature for Estimating Local Curvature in High Dimensional Data
• Batch-Size Independent Regret Bounds for Combinatorial Semi-Bandits with Probabilistically Triggered Arms or Independent Arms
• On Enforcing Better Conditioned Meta-Learning for Rapid Few-Shot Adaptation
• Efficient learning of nonlinear prediction models with time-series privileged information
• Training and Inference on Any-Order Autoregressive Models the Right Way
• SPD: Synergy Pattern Diversifying Oriented Unsupervised Multi-agent Reinforcement Learning
• GAPX: Generalized Autoregressive Paraphrase-Identification X
• CATER: Intellectual Property Protection on Text Generation APIs via Conditional Watermarks
• Reinforcement Learning with a Terminator
• Bringing Image Scene Structure to Video via Frame-Clip Consistency of Object Tokens
• Class-Dependent Label-Noise Learning with Cycle-Consistency Regularization
• CAGroup3D: Class-Aware Grouping for 3D Object Detection on Point Clouds
• Sparse Interaction Additive Networks via Feature Interaction Detection and Sparse Selection
• Object-Category Aware Reinforcement Learning
• Decision-Focused Learning without Decision-Making: Learning Locally Optimized Decision Losses
• Universally Expressive Communication in Multi-Agent Reinforcement Learning
• Are GANs overkill for NLP?
• Simple Mechanisms for Welfare Maximization in Rich Advertising Auctions
• Scalable Interpretability via Polynomials
• NOTE: Robust Continual Test-time Adaptation Against Temporal Correlation
• Learning Audio-Visual Dynamics Using Scene Graphs for Audio Source Separation
• Symmetry Teleportation for Accelerated Optimization
• The Nature of Temporal Difference Errors in Multi-step Distributional Reinforcement Learning
• Truncated proposals for scalable and hassle-free simulation-based inference
• Large-Scale Retrieval for Reinforcement Learning
• Decoupled Self-supervised Learning for Graphs
• In Differential Privacy, There is Truth: on Vote-Histogram Leakage in Ensemble Private Learning
• Handcrafted Backdoors in Deep Neural Networks
• Structuring Representations Using Group Invariants
• A sharp NMF result with applications in network modeling
• Improving Policy Learning via Language Dynamics Distillation
• Pure Transformers are Powerful Graph Learners
• Contextual Squeeze-and-Excitation for Efficient Few-Shot Image Classification
• Few-shot Image Generation via Adaptation-Aware Kernel Modulation
• Towards Understanding Grokking: An Effective Theory of Representation Learning
• Online Agnostic Multiclass Boosting
• Adversarial Unlearning: Reducing Confidence Along Adversarial Directions
• Robust Imitation via Mirror Descent Inverse Reinforcement Learning
• HyperMiner: Topic Taxonomy Mining with Hyperbolic Embedding
• Oracle-Efficient Online Learning for Smoothed Adversaries
• Multiclass Learnability Beyond the PAC Framework: Universal Rates and Partial Concept Classes
• Lower Bounds on Randomly Preconditioned Lasso via Robust Sparse Designs
• Learning from Distributed Users in Contextual Linear Bandits Without Sharing the Context
• Accelerating Sparse Convolution with Column Vector-Wise Sparsity
• Fast Instrument Learning with Faster Rates
• LTMD: Learning Improvement of Spiking Neural Networks with Learnable Thresholding Neurons and Moderate Dropout
• Improving Neural Ordinary Differential Equations with Nesterov's Accelerated Gradient Method
• Learning Neural Set Functions Under the Optimal Subset Oracle
• Guaranteed Conservation of Momentum for Learning Particle-based Fluid Dynamics
• Universality of Group Convolutional Neural Networks Based on Ridgelet Analysis on Groups
• On the Spectral Bias of Convolutional Neural Tangent and Gaussian Process Kernels
• Beyond L1: Faster and Better Sparse Models with skglm
• Improving GANs with A Dynamic Discriminator
• Streaming Radiance Fields for 3D Video Synthesis
• On the non-universality of deep learning: quantifying the cost of symmetry
• GraB: Finding Provably Better Data Permutations than Random Reshuffling
• Enhancing Safe Exploration Using Safety State Augmentation
• Robust Binary Models by Pruning Randomly-initialized Networks
• Optimal and Adaptive Monteiro-Svaiter Acceleration
• Reinforcement Learning with Logarithmic Regret and Policy Switches
• HYPRO: A Hybridly Normalized Probabilistic Model for Long-Horizon Prediction of Event Sequences
• Temporally-Consistent Survival Analysis
• Data-IQ: Characterizing subgroups with heterogeneous outcomes in tabular data
• Learning and Covering Sums of Independent Random Variables with Unbounded Support
• Learning to Discover and Detect Objects
• UViM: A Unified Modeling Approach for Vision with Learned Guiding Codes
• BOME! Bilevel Optimization Made Easy: A Simple First-Order Approach
• DeepFoids: Adaptive Bio-Inspired Fish Simulation with Deep Reinforcement Learning
• Improving Intrinsic Exploration with Language Abstractions
• MACE: Higher Order Equivariant Message Passing Neural Networks for Fast and Accurate Force Fields
• Monte Carlo Augmented Actor-Critic for Sparse Reward Deep Reinforcement Learning from Suboptimal Demonstrations
• ShapeCrafter: A Recursive Text-Conditioned 3D Shape Generation Model
• Non-identifiability and the Blessings of Misspecification in Models of Molecular Fitness
• VoiceBlock: Privacy through Real-Time Adversarial Attacks with Audio-to-Audio Models
• Recall Distortion in Neural Network Pruning and the Undecayed Pruning Algorithm
• Outlier-Robust Sparse Mean Estimation for Heavy-Tailed Distributions
• Invariance-Aware Randomized Smoothing Certificates
• Beyond the Return: Off-policy Function Estimation under User-specified Error-measuring Distributions
• On the Statistical Efficiency of Reward-Free Exploration in Non-Linear RL
• Energy-Based Contrastive Learning of Visual Representations
• Identifying good directions to escape the NTK regime and efficiently learn low-degree plus sparse polynomials
• Deep Surrogate Assisted Generation of Environments
• Hierarchical Lattice Layer for Partially Monotone Neural Networks
• SemiFL: Semi-Supervised Federated Learning for Unlabeled Clients with Alternate Training
• Self-Similarity Priors: Neural Collages as Differentiable Fractal Representations
• Sample Constrained Treatment Effect Estimation
• What's the Harm? Sharp Bounds on the Fraction Negatively Affected by Treatment
• Empirical Phase Diagram for Three-layer Neural Networks with Infinite Width
• FiLM-Ensemble: Probabilistic Deep Learning via Feature-wise Linear Modulation
• Maximum Likelihood Training of Implicit Nonlinear Diffusion Model
• Single Loop Gaussian Homotopy Method for Non-convex Optimization
• GAL: Gradient Assisted Learning for Decentralized Multi-Organization Collaborations
• CodeRL: Mastering Code Generation through Pretrained Models and Deep Reinforcement Learning
• Inducing Equilibria via Incentives: Simultaneous Design-and-Play Ensures Global Convergence
• Reinforcement Learning with Neural Radiance Fields
• Multi-agent Performative Prediction with Greedy Deployment and Consensus Seeking Agents
• A Differentially Private Linear-Time fPTAS for the Minimum Enclosing Ball Problem
• Debiased Causal Tree: Heterogeneous Treatment Effects Estimation with Unmeasured Confounding
• Assistive Teaching of Motor Control Tasks to Humans
• Learning interacting dynamical systems with latent Gaussian process ODEs
• Provably expressive temporal graph networks
• A Universal Error Measure for Input Predictions Applied to Online Graph Problems
• On the difficulty of learning chaotic dynamics with RNNs
• Learning on the Edge: Online Learning with Stochastic Feedback Graphs
• Meta-ticket: Finding optimal subnetworks for few-shot learning within randomly initialized neural networks
• Adjoint-aided inference of Gaussian process driven differential equations
• Subquadratic Kronecker Regression with Applications to Tensor Decomposition
• Post-hoc estimators for learning to defer to an expert
• Polynomial-Time Optimal Equilibria with a Mediator in Extensive-Form Games
• Contrastive and Non-Contrastive Self-Supervised Learning Recover Global and Local Spectral Embedding Methods
• Discovered Policy Optimisation
• Decomposable Non-Smooth Convex Optimization with Nearly-Linear Gradient Oracle Complexity
• SPD domain-specific batch normalization to crack interpretable unsupervised domain adaptation in EEG
• Convexity Certificates from Hessians
• Holomorphic Equilibrium Propagation Computes Exact Gradients Through Finite Size Oscillations
• Log-Linear-Time Gaussian Processes Using Binary Tree Kernels
• Indicators of Attack Failure: Debugging and Improving Optimization of Adversarial Examples
• Continual Learning In Environments With Polynomial Mixing Times
• VisFIS: Visual Feature Importance Supervision with Right-for-the-Right-Reason Objectives
• Algorithms and Hardness for Learning Linear Thresholds from Label Proportions
• Enhanced Meta Reinforcement Learning via Demonstrations in Sparse Reward Environments
• Make Some Noise: Reliable and Efficient Single-Step Adversarial Training
• Transformer Memory as a Differentiable Search Index
• (De-)Randomized Smoothing for Decision Stump Ensembles
• Global Normalization for Streaming Speech Recognition in a Modular Framework
• Theoretically Better and Numerically Faster Distributed Optimization with Smoothness-Aware Quantization Techniques
• Learning Tractable Probabilistic Models from Inconsistent Local Estimates
• List-Decodable Sparse Mean Estimation via Difference-of-Pairs Filtering
• Normalizing Flows for Knockoff-free Controlled Feature Selection
• Debiased Machine Learning without Sample-Splitting for Stable Estimators
• Explicable Policy Search
• Robustness to Unbounded Smoothness of Generalized SignSGD
• Subgame Solving in Adversarial Team Games
• Autoregressive Perturbations for Data Poisoning
• Trust Region Policy Optimization with Optimal Transport Discrepancies: Duality and Algorithm for Continuous Actions
• Statistical Learning and Inverse Problems: A Stochastic Gradient Approach
• TPU-KNN: K Nearest Neighbor Search at Peak FLOP/s
• Self-Aware Personalized Federated Learning
• Unsupervised Visual Representation Learning via Mutual Information Regularized Assignment
• LiteTransformerSearch: Training-free Neural Architecture Search for Efficient Language Models
• Nonstationary Dual Averaging and Online Fair Allocation
• Leveraging Inter-Layer Dependency for Post -Training Quantization
• FOF: Learning Fourier Occupancy Field for Monocular Real-time Human Reconstruction
• Learning Expressive Meta-Representations with Mixture of Expert Neural Processes
• REVIVE: Regional Visual Representation Matters in Knowledge-Based Visual Question Answering
• Online Neural Sequence Detection with Hierarchical Dirichlet Point Process
• Exploring Figure-Ground Assignment Mechanism in Perceptual Organization
• DTG-SSOD: Dense Teacher Guidance for Semi-Supervised Object Detection
• Deliberated Domain Bridging for Domain Adaptive Semantic Segmentation
• Dual-Curriculum Contrastive Multi-Instance Learning for Cancer Prognosis Analysis with Whole Slide Images
• BadPrompt: Backdoor Attacks on Continuous Prompts
• Geodesic Self-Attention for 3D Point Clouds
• Learning Enhanced Representation for Tabular Data via Neighborhood Propagation
• Spectrum Random Masking for Generalization in Image-based Reinforcement Learning
• 3DB: A Framework for Debugging Computer Vision Models
• High-dimensional limit theorems for SGD: Effective dynamics and critical scaling
• Provable Generalization of Overparameterized Meta-learning Trained with SGD
• MinVIS: A Minimal Video Instance Segmentation Framework without Video-based Training
• Efficient Meta Reinforcement Learning for Preference-based Fast Adaptation
• Reinforced Genetic Algorithm for Structure-based Drug Design
• Motion Transformer with Global Intention Localization and Local Movement Refinement
• Deep Fourier Up-Sampling
• FR: Folded Rationalization with a Unified Encoder
• Measures of Information Reflect Memorization Patterns
• Trading off Image Quality for Robustness is not Necessary with Regularized Deterministic Autoencoders
• CASA: Category-agnostic Skeletal Animal Reconstruction
• Learning Energy Networks with Generalized Fenchel-Young Losses
• Regularized Gradient Descent Ascent for Two-Player Zero-Sum Markov Games
• Rethinking Image Restoration for Object Detection
• GBA: A Tuning-free Approach to Switch between Synchronous and Asynchronous Training for Recommendation Models
• Modeling Human Exploration Through Resource-Rational Reinforcement Learning
• SignRFF: Sign Random Fourier Features
• Gradient Estimation with Discrete Stein Operators
• Provably Efficient Reinforcement Learning in Partially Observable Dynamical Systems
• Single-phase deep learning in cortico-cortical networks
• GraphQNTK: Quantum Neural Tangent Kernel for Graph Data
• BiMLP: Compact Binary Architectures for Vision Multi-Layer Perceptrons
• Sampling in Constrained Domains with Orthogonal-Space Variational Gradient Descent
• Exploring the Limits of Domain-Adaptive Training for Detoxifying Large-Scale Language Models
• LasUIE: Unifying Information Extraction with Latent Adaptive Structure-aware Generative Language Model
• An $\alpha$-regret analysis of Adversarial Bilateral Trade
• Intrinsic dimensionality estimation using Normalizing Flows
• Supervised Training of Conditional Monge Maps
• Drawing out of Distribution with Neuro-Symbolic Generative Models
• Sketching based Representations for Robust Image Classification with Provable Guarantees
• Learning low-dimensional generalizable natural features from retina using a U-net
• Data Augmentation for Compositional Data: Advancing Predictive Models of the Microbiome
• VisCo Grids: Surface Reconstruction with Viscosity and Coarea Grids
• Synergy-of-Experts: Collaborate to Improve Adversarial Robustness
• Neural Matching Fields: Implicit Representation of Matching Fields for Visual Correspondence
• Fast Bayesian Estimation of Point Process Intensity as Function of Covariates
• MOVE: Unsupervised Movable Object Segmentation and Detection
• Not All Bits have Equal Value: Heterogeneous Precisions via Trainable Noise
• Differentially Private Learning Needs Hidden State (Or Much Faster Convergence)
• Training Spiking Neural Networks with Local Tandem Learning
• Unsupervised Skill Discovery via Recurrent Skill Training
• Interpreting Operation Selection in Differentiable Architecture Search: A Perspective from Influence-Directed Explanations
• Fair Rank Aggregation
• Optimal Gradient Sliding and its Application to Optimal Distributed Optimization Under Similarity
• Contact-aware Human Motion Forecasting
• Non-rigid Point Cloud Registration with Neural Deformation Pyramid
• Make an Omelette with Breaking Eggs: Zero-Shot Learning for Novel Attribute Synthesis
• The First Optimal Algorithm for Smooth and Strongly-Convex-Strongly-Concave Minimax Optimization
• Towards Reasonable Budget Allocation in Untargeted Graph Structure Attacks via Gradient Debias
• Stability and Generalization of Kernel Clustering: from Single Kernel to Multiple Kernel
• Few-shot Relational Reasoning via Connection Subgraph Pretraining
• Alleviating Adversarial Attacks on Variational Autoencoders with MCMC
• Coreset for Line-Sets Clustering
• Fast Stochastic Composite Minimization and an Accelerated Frank-Wolfe Algorithm under Parallelization
• FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness
• HF-NeuS: Improved Surface Reconstruction Using High-Frequency Details
• On the Effectiveness of Fine-tuning Versus Meta-reinforcement Learning
• Spatial Pruned Sparse Convolution for Efficient 3D Object Detection
• Byzantine Spectral Ranking
• What Can the Neural Tangent Kernel Tell Us About Adversarial Robustness?
• On Translation and Reconstruction Guarantees of the Cycle-Consistent Generative Adversarial Networks
• Evaluated CMI Bounds for Meta Learning: Tightness and Expressiveness
• SnAKe: Bayesian Optimization with Pathwise Exploration
• Random Rank: The One and Only Strategyproof and Proportionally Fair Randomized Facility Location Mechanism
• Resolving the data ambiguity for periodic crystals
• CroCo: Self-Supervised Pre-training for 3D Vision Tasks by Cross-View Completion
• Coresets for Vertical Federated Learning: Regularized Linear Regression and $K$-Means Clustering
• Learning Predictions for Algorithms with Predictions
• Hyperparameter Sensitivity in Deep Outlier Detection: Analysis and a Scalable Hyper-Ensemble Solution
• DASCO: Dual-Generator Adversarial Support Constrained Offline Reinforcement Learning
• Exploring through Random Curiosity with General Value Functions
• Equivariant Networks for Zero-Shot Coordination
• A PAC-Bayesian Generalization Bound for Equivariant Networks
• Split-kl and PAC-Bayes-split-kl Inequalities for Ternary Random Variables
• Pareto Set Learning for Expensive Multi-Objective Optimization
• Formalizing Consistency and Coherence of Representation Learning
• Compositional generalization through abstract representations in human and artificial neural networks
• The Sample Complexity of One-Hidden-Layer Neural Networks
• Diffusion Visual Counterfactual Explanations
• Finding Optimal Arms in Non-stochastic Combinatorial Bandits with Semi-bandit Feedback and Finite Budget
• Pessimism for Offline Linear Contextual Bandits using $\ell_p$ Confidence Sets
• Assaying Out-Of-Distribution Generalization in Transfer Learning
• What are the best Systems? New Perspectives on NLP Benchmarking
• Clipped Stochastic Methods for Variational Inequalities with Heavy-Tailed Noise
• Hardness in Markov Decision Processes: Theory and Practice
• Generalization Error Bounds on Deep Learning with Markov Datasets
• Information-Theoretic Safe Exploration with Gaussian Processes
• M³ViT: Mixture-of-Experts Vision Transformer for Efficient Multi-task Learning with Model-Accelerator Co-design
• HierSpeech: Bridging the Gap between Text and Speech by Hierarchical Variational Inference using Self-supervised Representations for Speech Synthesis
• [Re] Replication Study of "Fairness and Bias in Online Selection"
• Triangulation candidates for Bayesian optimization
• Non-asymptotic and Accurate Learning of Nonlinear Dynamical Systems
• Washing The Unwashable : On The (Im)possibility of Fairwashing Detection
• No-regret learning in games with noisy feedback: Faster rates and adaptivity via learning rate separation
• Adaptive Data Debiasing through Bounded Exploration
• Positive-Unlabeled Learning using Random Forests via Recursive Greedy Risk Minimization
• Toward a realistic model of speech processing in the brain with self-supervised learning
• TabNAS: Rejection Sampling for Neural Architecture Search on Tabular Datasets
• CLIPDraw: Exploring Text-to-Drawing Synthesis through Language-Image Encoders
• Attention-based Neural Cellular Automata
• Sparse Additive Gaussian Process Regression
• Attraction-Repulsion Spectrum in Neighbor Embeddings
• Online Nonnegative CP-dictionary Learning for Markovian Data
• Decimated Framelet System on Graphs and Fast G-Framelet Transforms
• Multi-Agent Multi-Armed Bandits with Limited Communication
• Accelerated Zeroth-Order and First-Order Momentum Methods from Mini to Minimax Optimization
• Optimality and Stability in Non-Convex Smooth Games
• Deep Limits and a Cut-Off Phenomenon for Neural Networks
• Robust and scalable manifold learning via landmark diffusion for long-term medical signal processing
• [Re] Differentiable Spatial Planning using Transformers
• All You Need is a Good Functional Prior for Bayesian Deep Learning
• Recovery and Generalization in Over-Realized Dictionary Learning
• Truncated Emphatic Temporal Difference Methods for Prediction and Control
• [Re] AdaBelief Optimizer: Adapting Stepsizes by the Belief in Observed Gradients
• When is the Convergence Time of Langevin Algorithms Dimension Independent? A Composite Optimization Viewpoint
• [Re] Replication study of 'Data-Driven Methods for Balancing Fairness and Efficiency in Ride-Pooling'
• Learning Operators with Coupled Attention
• [Re] Solving Phase Retrieval With a Learned Reference
• [Re] Explaining in Style: Training a GAN to explain a classifier in StyleSpace
• DeepInteraction: 3D Object Detection via Modality Interaction
• Mix and Reason: Reasoning over Semantic Topology with Data Mixing for Domain Generalization
• SGAM: Building a Virtual 3D World through Simultaneous Generation and Mapping
• RLIP: Relational Language-Image Pre-training for Human-Object Interaction Detection
• Dense Interspecies Face Embedding
• Orthogonal Transformer: An Efficient Vision Transformer Backbone with Token Orthogonalization
• UMIX: Improving Importance Weighting for Subpopulation Shift via Uncertainty-Aware Mixup
• Rethinking Individual Global Max in Cooperative Multi-Agent Reinforcement Learning
• Resource-Adaptive Federated Learning with All-In-One Neural Composition
• One Model to Edit Them All: Free-Form Text-Driven Image Manipulation with Semantic Modulations
• Weakly-Supervised Multi-Granularity Map Learning for Vision-and-Language Navigation
• On the Robustness of Deep Clustering Models: Adversarial Attacks and Defenses
• Uncoupled Learning Dynamics with $O(\log T)$ Swap Regret in Multiplayer Games
• Weak-shot Semantic Segmentation via Dual Similarity Transfer
• Effective Backdoor Defense by Exploiting Sensitivity of Poisoned Samples
• Obj2Seq: Formatting Objects as Sequences with Class Prompt for Visual Tasks
• OTKGE: Multi-modal Knowledge Graph Embeddings via Optimal Transport
• Positively Weighted Kernel Quadrature via Subsampling
• LASSIE: Learning Articulated Shapes from Sparse Image Ensemble via 3D Part Discovery
• A Kernelised Stein Statistic for Assessing Implicit Generative Models
• E-MAPP: Efficient Multi-Agent Reinforcement Learning with Parallel Program Guidance
• EpiGRAF: Rethinking training of 3D GANs
• Bridging the Gap Between Vision Transformers and Convolutional Neural Networks on Small Datasets
• Optimal Efficiency-Envy Trade-Off via Optimal Transport
• Generating Long Videos of Dynamic Scenes
• Private Synthetic Data for Multitask Learning and Marginal Queries
• Graph Self-supervised Learning with Accurate Discrepancy Learning
• Independence Testing for Bounded Degree Bayesian Networks
• Tikhonov Regularization is Optimal Transport Robust under Martingale Constraints
• ZeroC: A Neuro-Symbolic Model for Zero-shot Concept Recognition and Acquisition at Inference Time
• SAMURAI: Shape And Material from Unconstrained Real-world Arbitrary Image collections
• Outlier Suppression: Pushing the Limit of Low-bit Transformer Language Models
• Bayesian Persuasion for Algorithmic Recourse
• Deep Hierarchical Planning from Pixels
• Noise Attention Learning: Enhancing Noise Robustness by Gradient Scaling
• Neural Basis Models for Interpretability
• Hierarchical classification at multiple operating points
• Information-Theoretic GAN Compression with Variational Energy-based Model
• Redistribution of Weights and Activations for AdderNet Quantization
• Deep invariant networks with differentiable augmentation layers
• Convergence beyond the over-parameterized regime using Rayleigh quotients
• Robust $\phi$-Divergence MDPs
• ToDD: Topological Compound Fingerprinting in Computer-Aided Drug Discovery
• On Privacy and Personalization in Cross-Silo Federated Learning
• Differentially Private Covariance Revisited
• Learning Graph-embedded Key-event Back-tracing for Object Tracking in Event Clouds
• Distributional Convergence of the Sliced Wasserstein Process
• Homomorphic Matrix Completion
• Transfer Learning on Heterogeneous Feature Spaces for Treatment Effects Estimation
• On the Identifiability of Nonlinear ICA: Sparsity and Beyond
• Museformer: Transformer with Fine- and Coarse-Grained Attention for Music Generation
• Towards Diverse and Faithful One-shot Adaption of Generative Adversarial Networks
• Dance of SNN and ANN: Solving binding problem by combining spike timing and reconstructive attention
• Efficient Sampling on Riemannian Manifolds via Langevin MCMC
• ATD: Augmenting CP Tensor Decomposition by Self Supervision
• Imitating Past Successes can be Very Suboptimal
• RKHS-SHAP: Shapley Values for Kernel Methods
• SAPD+: An Accelerated Stochastic Method for Nonconvex-Concave Minimax Problems
• On Scalable Testing of Samplers
• Markovian Interference in Experiments
• DP-PCA: Statistically Optimal and Differentially Private PCA
• Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learning
• Continual learning: a feature extraction formalization, an efficient algorithm, and fundamental obstructions
• Functional Ensemble Distillation
• Self-explaining deep models with logic rule reasoning
• Benign Underfitting of Stochastic Gradient Descent
• Modeling the Machine Learning Multiverse
• Stability Analysis and Generalization Bounds of Adversarial Training
• Exact Shape Correspondence via 2D graph convolution
• A Theoretical Understanding of Gradient Bias in Meta-Reinforcement Learning
• How and Why to Manipulate Your Own Agent: On the Incentives of Users of Learning Agents
• MissDAG: Causal Discovery in the Presence of Missing Data with Continuous Additive Noise Models
• Phase diagram of Stochastic Gradient Descent in high-dimensional two-layer neural networks
• Stochastic Online Learning with Feedback Graphs: Finite-Time and Asymptotic Optimality
• Spectral Bias Outside the Training Set for Deep Networks in the Kernel Regime
• First Hitting Diffusion Models for Generating Manifold, Graph and Categorical Data
• Universal Rates for Interactive Learning
• DGD^2: A Linearly Convergent Distributed Algorithm For High-dimensional Statistical Recovery
• Single-Stage Visual Relationship Learning using Conditional Queries
• Pruning has a disparate impact on model accuracy
• Teacher Forcing Recovers Reward Functions for Text Generation
• Model-based RL with Optimistic Posterior Sampling: Structural Conditions and Sample Complexity
• Optimal Dynamic Regret in LQR Control
• Generalization Gap in Amortized Inference
• Near-Optimal Private and Scalable $k$-Clustering
• Language Models with Image Descriptors are Strong Few-Shot Video-Language Learners
• Hedging as Reward Augmentation in Probabilistic Graphical Models
• Training Subset Selection for Weak Supervision
• Online Reinforcement Learning for Mixed Policy Scopes
• Branch & Learn for Recursively and Iteratively Solvable Problems in Predict+Optimize
• Your Out-of-Distribution Detection Method is Not Robust!
• An efficient graph generative model for navigating ultra-large combinatorial synthesis libraries
• Communication-Efficient Topologies for Decentralized Learning with $O(1)$ Consensus Rate
• Rethinking the Reverse-engineering of Trojan Triggers
• Decentralized, Communication- and Coordination-free Learning in Structured Matching Markets
• On the Epistemic Limits of Personalized Prediction
• Learning to Mitigate AI Collusion on Economic Platforms
• STNDT: Modeling Neural Population Activity with Spatiotemporal Transformers
• Masked Autoencoding for Scalable and Generalizable Decision Making
• Computationally Efficient Horizon-Free Reinforcement Learning for Linear Mixture MDPs
• DiSC: Differential Spectral Clustering of Features
• Personalized Online Federated Learning with Multiple Kernels
• Patching open-vocabulary models by interpolating weights
• Concrete Score Matching: Generalized Score Matching for Discrete Data
• LBD: Decouple Relevance and Observation for Individual-Level Unbiased Learning to Rank
• Palm up: Playing in the Latent Manifold for Unsupervised Pretraining
• Focal Modulation Networks
• S2P: State-conditioned Image Synthesis for Data Augmentation in Offline Reinforcement Learning
• Exploitability Minimization in Games and Beyond
• FeLMi : Few shot Learning with hard Mixup
• The First Optimal Acceleration of High-Order Methods in Smooth Convex Optimization
• On Optimal Learning Under Targeted Data Poisoning
• The computational and learning benefits of Daleian neural networks
• Support Recovery in Sparse PCA with Incomplete Data
• Missing Data Imputation and Acquisition with Deep Hierarchical Models and Hamiltonian Monte Carlo
• Private Isotonic Regression
• Do Residual Neural Networks discretize Neural Ordinary Differential Equations?
• Continuous MDP Homomorphisms and Homomorphic Policy Gradient
• Exponentially Improving the Complexity of Simulating the Weisfeiler-Lehman Test with Graph Neural Networks
• Look where you look! Saliency-guided Q-networks for generalization in visual Reinforcement Learning
• Constrained GPI for Zero-Shot Transfer in Reinforcement Learning
• Predicting Cellular Responses to Novel Drug Perturbations at a Single-Cell Resolution
• A Boosting Approach to Reinforcement Learning
• DataMUX: Data Multiplexing for Neural Networks
• Are Defenses for Graph Neural Networks Robust?
• Adversarial Robustness is at Odds with Lazy Training
• Robust Reinforcement Learning using Offline Data
• Lifting the Information Ratio: An Information-Theoretic Analysis of Thompson Sampling for Contextual Bandits
• Fine-tuning language models to find agreement among humans with diverse preferences
• Tsetlin Machine for Solving Contextual Bandit Problems
• Multi-Class $H$-Consistency Bounds
• Statistical, Robustness, and Computational Guarantees for Sliced Wasserstein Distances
• Lifting Weak Supervision To Structured Prediction
• Learning Concept Credible Models for Mitigating Shortcuts
• LST: Ladder Side-Tuning for Parameter and Memory Efficient Transfer Learning
• Disentangling Transfer in Continual Reinforcement Learning
• Distinguishing discrete and continuous behavioral variability using warped autoregressive HMMs
• Off-Team Learning
• LAMP: Extracting Text from Gradients with Language Model Priors
• 4D Unsupervised Object Discovery
• Bayesian subset selection and variable importance for interpretable prediction and classification
• Unifying Voxel-based Representation with Transformer for 3D Object Detection
• Multi-Scale Adaptive Network for Single Image Denoising
• Improving Out-of-Distribution Generalization by Adversarial Training with Structured Priors
• Efficient Graph Similarity Computation with Alignment Regularization
• Revisiting Realistic Test-Time Training: Sequential Inference and Adaptation by Anchored Clustering
• Falconn++: A Locality-sensitive Filtering Approach for Approximate Nearest Neighbor Search
• Natural image synthesis for the retina with variational information bottleneck representation
• A Lower Bound of Hash Codes' Performance
• I2Q: A Fully Decentralized Q-Learning Algorithm
• Shadow Knowledge Distillation: Bridging Offline and Online Knowledge Transfer
• A Differentiable Semantic Metric Approximation in Probabilistic Embedding for Cross-Modal Retrieval
• Heterogeneous Skill Learning for Multi-agent Tasks
• Are You Stealing My Model? Sample Correlation for Fingerprinting Deep Neural Networks
• Model-Based Opponent Modeling
• When Adversarial Training Meets Vision Transformers: Recipes from Training to Architecture
• Learning Invariant Graph Representations for Out-of-Distribution Generalization
• Learning from Future: A Novel Self-Training Framework for Semantic Segmentation
• Geo-Neus: Geometry-Consistent Neural Implicit Surfaces Learning for Multi-view Reconstruction
• Wasserstein Iterative Networks for Barycenter Estimation
• Personalized Federated Learning towards Communication Efficiency, Robustness and Fairness
• Analyzing Sharpness along GD Trajectory: Progressive Sharpening and Edge of Stability
• Rashomon Capacity: A Metric for Predictive Multiplicity in Classification
• Pre-trained Adversarial Perturbations
• Conformal Frequency Estimation with Sketched Data
• Convergent Representations of Computer Programs in Human and Artificial Neural Networks
• tntorch: Tensor Network Learning with PyTorch
• Monte Carlo Tree Search based Variable Selection for High Dimensional Bayesian Optimization
• Joint Entropy Search for Multi-Objective Bayesian Optimization
• Embracing Consistency: A One-Stage Approach for Spatio-Temporal Video Grounding
• A Closer Look at the Adversarial Robustness of Deep Equilibrium Models
• Language Conditioned Spatial Relation Reasoning for 3D Object Grounding
• [Re] Projection-based Algorithm for Updating the TruncatedSVD of Evolving Matrices
• Unsupervised Point Cloud Completion and Segmentation by Generative Adversarial Autoencoding Network
• Audio-Driven Co-Speech Gesture Video Generation
• Dynamic Graph Neural Networks Under Spatio-Temporal Distribution Shift
• On the role of overparameterization in off-policy Temporal Difference learning with linear function approximation
• Quantized Training of Gradient Boosting Decision Trees
• InterpretDL: Explaining Deep Models in PaddlePaddle
• Efficient Change-Point Detection for Tackling Piecewise-Stationary Bandits
• A Unified Statistical Learning Model for Rankings and Scores with Application to Grant Panel Review
• LSAR: Efficient Leverage Score Sampling Algorithm for the Analysis of Big Time Series Data
• D-GCCA: Decomposition-based Generalized Canonical Correlation Analysis for Multi-view High-dimensional Data
• Supervised Dimensionality Reduction and Visualization using Centroid-Encoder
• Foolish Crowds Support Benign Overfitting
• Rethinking Nonlinear Instrumental Variable Models through Prediction Validity
• [Re] Reproduction and Extension of "Queens are Powerful too: Mitigating Gender Bias in Dialogue Generation"
• [Re] An Implementation of Fair Robust Learning
• GAMA: Generative Adversarial Multi-Object Scene Attacks
• Point-M2AE: Multi-scale Masked Autoencoders for Hierarchical Point Cloud Pre-training
• An Empirical Study on Disentanglement of Negative-free Contrastive Learning
• Social-Inverse: Inverse Decision-making of Social Contagion Management with Task Migrations
• Bayesian Risk Markov Decision Processes
• SHINE: SubHypergraph Inductive Neural nEtwork
• Singular Value Fine-tuning: Few-shot Segmentation requires Few-parameters Fine-tuning
• Non-deep Networks
• Feature-Proxy Transformer for Few-Shot Segmentation
• Estimating and Explaining Model Performance When Both Covariates and Labels Shift
• The alignment property of SGD noise and how it helps select flat minima: A stability analysis
• Thompson Sampling Efficiently Learns to Control Diffusion Processes
• Fair and Efficient Allocations Without Obvious Manipulations
• Fuzzy Learning Machine
• ASPiRe: Adaptive Skill Priors for Reinforcement Learning
• Sound and Complete Causal Identification with Latent Variables Given Local Background Knowledge
• Explainability Via Causal Self-Talk
• ALIFE: Adaptive Logit Regularizer and Feature Replay for Incremental Semantic Segmentation
• Causal Identification under Markov equivalence: Calculus, Algorithm, and Completeness
• Random Normalization Aggregation for Adversarial Defense
• Size and depth of monotone neural networks: interpolation and approximation
• Spartan: Differentiable Sparsity via Regularized Transportation
• On Gap-dependent Bounds for Offline Reinforcement Learning
• Revisiting Sliced Wasserstein on Images: From Vectorization to Convolution
• Accelerated Projected Gradient Algorithms for Sparsity Constrained Optimization Problems
• Context-Based Dynamic Pricing with Partially Linear Demand Model
• S-PIFu: Integrating Parametric Human Models with PIFu for Single-view Clothed Human Reconstruction
• Action-modulated midbrain dopamine activity arises from distributed control policies
• NUWA-Infinity: Autoregressive over Autoregressive Generation for Infinite Visual Synthesis
• Bezier Gaussian Processes for Tall and Wide Data
• Minimax Regret for Cascading Bandits
• Pre-activation Distributions Expose Backdoor Neurons
• Posterior Matching for Arbitrary Conditioning
• Alternating Mirror Descent for Constrained Min-Max Games
• Early Stage Convergence and Global Convergence of Training Mildly Parameterized Neural Networks
• Foundation Posteriors for Approximate Probabilistic Inference
• Entropy-Driven Mixed-Precision Quantization for Deep Network Design
• Denoising Diffusion Restoration Models
• Syndicated Bandits: A Framework for Auto Tuning Hyper-parameters in Contextual Bandit Algorithms
• A Combinatorial Perspective on the Optimization of Shallow ReLU Networks
• A Lagrangian Duality Approach to Active Learning
• A Statistical Online Inference Approach in Averaged Stochastic Approximation
• Theoretical analysis of deep neural networks for temporally dependent observations
• On the Complexity of Adversarial Decision Making
• Scalable Distributional Robustness in a Class of Non-Convex Optimization with Guarantees
• Uncovering the Structural Fairness in Graph Contrastive Learning
• ULNeF: Untangled Layered Neural Fields for Mix-and-Match Virtual Try-On
• Tree ensemble kernels for Bayesian optimization with known constraints over mixed-feature spaces
• Better Uncertainty Calibration via Proper Scores for Classification and Beyond
• Augmenting Online Algorithms with $\varepsilon$-Accurate Predictions
• Asymptotics of $\ell_2$ Regularized Network Embeddings
• Deep Counterfactual Estimation with Categorical Background Variables
• What is a Good Metric to Study Generalization of Minimax Learners?
• Non-Convex Bilevel Games with Critical Point Selection Maps
• The Franz-Parisi Criterion and Computational Trade-offs in High Dimensional Statistics
• Provable Defense against Backdoor Policies in Reinforcement Learning
• IMED-RL: Regret optimal learning of ergodic Markov decision processes
• Leveraging the Hints: Adaptive Bidding in Repeated First-Price Auctions
• A Closer Look at Learned Optimization: Stability, Robustness, and Inductive Biases
• Distributed Learning of Finite Gaussian Mixtures
• [Re] Reproducibility Report: Contrastive Learning of Socially-aware Motion Representations
• [Re] GANSpace: Discovering Interpretable GAN Controls
• [Re] Reproducibility Study of “Counterfactual Generative Networks”
• [Re] Does Self-Supervision Always Improve Few-Shot Learning?
• On Kernelized Multi-Armed Bandits with Constraints
• A Nonconvex Framework for Structured Dynamic Covariance Recovery
• A Mean-Field Game Approach to Cloud Resource Management with Function Approximation
• CoupAlign: Coupling Word-Pixel with Sentence-Mask Alignments for Referring Image Segmentation
• Fast and Robust Rank Aggregation against Model Misspecification
• Poisson Flow Generative Models
• Boosting Out-of-distribution Detection with Typical Features
• Joint Estimation and Inference for Data Integration Problems based on Multiple Multi-layered Gaussian Graphical Models
• Conditional Diffusion Process for Inverse Halftoning
• Knowledge Distillation from A Stronger Teacher
• Dynamic pricing and assortment under a contextual MNL demand
• Temporal Effective Batch Normalization in Spiking Neural Networks
• The trade-offs of model size in large recommendation models : 100GB to 10MB Criteo-tb DLRM model
• Neural Transmitted Radiance Fields
• 🏘️ ProcTHOR: Large-Scale Embodied AI Using Procedural Generation
• [Re] Exacerbating Algorithmic Bias through Fairness Attacks
• Transfer Learning in Information Criteria-based Feature Selection
• Learning with little mixing
• Fairness-Aware PAC Learning from Corrupted Data
• Online Frank-Wolfe with Arbitrary Delays
• Online Allocation and Learning in the Presence of Strategic Agents
• Weisfeiler and Leman Go Walking: Random Walk Kernels Revisited
• Sufficient reductions in regression with mixed predictors
• Optimal Query Complexities for Dynamic Trace Estimation
• Understanding Benign Overfitting in Gradient-Based Meta Learning
• Generalised Mutual Information for Discriminative Clustering
• Prune and distill: similar reformatting of image information along rat visual cortex and deep neural networks
• ShuffleMixer: An Efficient ConvNet for Image Super-Resolution
• Minimax Optimal Fixed-Budget Best Arm Identification in Linear Bandits
• [Re] Strategic classification made practical: reproduction
• Communication-efficient distributed eigenspace estimation with arbitrary node failures
• Generalised Implicit Neural Representations
• AttCAT: Explaining Transformers via Attentive Class Activation Tokens
• Score-Based Generative Models Detect Manifolds
• Geodesic Graph Neural Network for Efficient Graph Representation Learning
• [Re] Nondeterminism and Instability in Neural Network Optimization
• Diffusion-based Molecule Generation with Informative Prior Bridges
• [Re] Learning Unknown from Correlations: Graph Neural Network for Inter-novel-protein Interaction Prediction
• The Importance of Being Correlated: Implications of Dependence in Joint Spectral Inference across Multiple Networks
• Zero-shot Transfer Learning within a Heterogeneous Graph via Knowledge Transfer Networks
• Provably sample-efficient RL with side information about latent dynamics
• Signal Processing for Implicit Neural Representations
• Embrace the Gap: VAEs Perform Independent Mechanism Analysis
• Exponential Separations in Symmetric Neural Networks
• [Re] Understanding Self-Supervised Learning Dynamics without Contrastive Pairs
• Transformers from an Optimization Perspective
• (f,Gamma)-Divergences: Interpolating between f-Divergences and Integral Probability Metrics
• PointNeXt: Revisiting PointNet++ with Improved Training and Scaling Strategies
• [Re] Transparent Object Tracking Benchmark
• Neural Lyapunov Control of Unknown Nonlinear Systems with Stability Guarantees
• [Re] Graph Edit Networks
• [Re] Reproduction Study of Variational Fair Clustering
• BR-SNIS: Bias Reduced Self-Normalized Importance Sampling
• Fair and Optimal Decision Trees: A Dynamic Programming Approach
• Brain Network Transformer
• Learning to Navigate Wikipedia by Taking Random Walks
• The Neural Testbed: Evaluating Joint Predictions
• A Bregman Learning Framework for Sparse Neural Networks
• [Re] Learning to count everything
• On the Double Descent of Random Features Models Trained with SGD

• Live on Sky
• Get Sky Sports
• Sky Mobile Apps
• Kick It Out
• Black Lives Matter
• British South Asians in Football

Transfer Centre

Mauricio Pochettino interested in talking to Man Utd if job becomes available - Paper Talk

Plus: Arsene Wenger is to press ahead with his proposal to bring in the most radical change to the offside law for more than 30 years; Mauricio Pochettino will be a leading candidate for the England job

Thursday 23 May 2024 07:25, UK

Please use Chrome browser for a more accessible video player

The top stories and transfer rumours from Thursday's newspapers...

Mauricio Pochettino is interested in talking to Manchester United if they decide to part ways with Erik ten Hag this summer.

Arsene Wenger is to press ahead with his proposal to bring in the most radical change to the offside law for more than 30 years after what he views as positive results from trials.

Transfer Centre LIVE!

• Sky Sports+ - more than 1,000 live EFL games next season at no extra cost
• All you need to know - Streaming Sky Sports with NOW
• Get Sky Sports on WhatsApp | Download the Sky Sports app

It will be a race against time for the independent football regulator to be approved by parliament after the prime minister, Rishi Sunak, announced a general election for July 4.

• Transfer Centre LIVE! Ipswich supremo Schwartz flies in for McKenna talks
• Papers: Chelsea to make £60m move for Palace winger Olise
• 'We go for next season' - is Ten Hag expecting Man Utd stay?
• Okolie blows away Rozanski to become two-weight world champion
• Arum: Fury anxious for Usyk rematch... then AJ at Wembley?
• Barcelona sack head coach Xavi with Flick primed to replace him
• Man Utd transfers - Fernandes: I don't want to leave
• 'Our best day of 2024!' - Hamilton buoyed by Mercedes' Monaco pace
• Ten Hag dismisses Man Utd exit talk | Pep sympathy over injuries
• McGregor to make UFC comeback in June
• Latest News

Eddie Howe says he has no injury concerns over Kieran Trippier after the Newcastle defender limped off early in a post-season friendly and was later seen with an ice pack on his ankle.

THE GUARDIAN

Stream Sky Sports on NOW

Download the Sky Sports app

Bayern Munich are close to ending their protracted search for a new manager by handing the vacancy to Burnley's Vincent Kompany.

Chelsea have intensified their search for Mauricio Pochettino's replacement by making checks on Leicester's Enzo Maresca, but candidates for the job believe that Ipswich's Kieran McKenna is the favourite to take over at Stamford Bridge.

The United States-based investment fund Oaktree Capital Management said on Wednesday it has become the new owner of Serie A champions Inter Milan after a missed $428m (£336m) payment from the club's Chinese holding company, Suning.

FIFA must be prepared to deny Saudi Arabia the right to host the 2034 men's World Cup if the kingdom fails to comply with human rights obligations, according to a new legal submission filed with the governing body.

The acting sports minister of Ukraine, Matviy Bidnyi, has told his country's athletes to keep a "cold head" and pay no attention to any provocation from their Russian counterparts at the Olympic Games this summer.

There is a clamour for a dedicated TV slot for women's football, according to a survey of thousands of fans carried out by the Football Supporters' Association.

DAILY TELEGRAPH

Mauricio Pochettino will be a leading candidate for the England job if he remains out of work and the post becomes available after Euro 2024.

Chelsea players were left in shock after Mauricio Pochettino left the club by mutual consent on Tuesday.

Dan Ashworth is attempting to force Newcastle United to let him out of his lengthy gardening leave early so that he can start work at Manchester United and will argue he was sacked from the club rather than resigning.

Bookmakers have slashed odds on Manchester City getting relegated next season as the club awaits its hearing for 115 alleged financial breaches.

An American consortium of Silicon Valley-based private investors have revived their hopes of a takeover at Sheffield United following their relegation from the Premier League.

The Culture Secretary says she is intensifying pressure on the Football Association to toughen up its transgender policy which she maintains is not "correct".

Stefan Ortega has revealed why he settled in so quickly at Manchester City - a visit to his local pub.

Ipswich have offered Kieran McKenna a blockbuster new contract to try and keep him at Portman Road.

Robert Lewandowski agreed to join Manchester United after a phone call with Sir Alex Ferguson... before Borussia Dortmund blocked the deal.

Mauricio Pochettino's exit from Chelsea has "ruined the feelgood factor" in the dressing room.

Mason Mount put himself in the frame for Saturday's FA Cup final by returning to training.

Manchester United's shambolic transfer strategy is as bad as ever under Erik ten Hag according to forensic analysis by Dutch experts BEBR.

New Liverpool boss Arne Slot is in pole position to sign Feyenoord's Justin Bijlow and head off a potential Kop keeper crisis but £64m striker Darwin Nunez could be on his way out.

DAILY MIRROR

Super-agent Jorge Mendes has put Porto boss Sergio Conceicao's name forward as a potential replacement for Mauricio Pochettino.

Gareth Southgate says his new-look England squad will have more "hunger" going into the Euros.

Newcastle are closing in on a deal to sign Fulham defender Tosin Adarabioyo on a free transfer.

Darwin Nunez has admitted that he was affected by 'negative comments' following his social media purge of Liverpool-related photos.

In-demand Ipswich manager Kieran McKenna is set to reject the offer of a new deal form the club, according to a report.

The FA and UEFA have vowed to crack down on online ticket touts after it emerged some seats for next Saturday's Champions League final at Wembley have been put on sale for as much as £134,000.

Wimbledon fans may get the chance to bid goodbye to Rafael Nadal as the Spaniard's name appeared on the tournament entry list.

Arsenal are in the market for Feyenoord goalkeeper Justin Bijlow with Mikel Arteta wanting a new No.2 as Aaron Ramsdale is expected to leave.

DAILY RECORD

Hibs will draw up a managerial shortlist in the coming days as they close in on appointing Nick Montgomery's successor.

Celtic have reportedly entered the race to sign Shakhtar Donetsk left-back Irakli Azarovi.

Correctly predict six scorelines for a chance to win £250,000 for free. Entries by 3pm Saturday.