Publications

A Theoretical Justification for Asymmetric Actor-Critic Algorithms
Gaspard Lambrechts
Damien Ernst
Aditya Mahajan
2025-05-01
ICML.cc/2025/Conference (poster)
openreview.net
openreview.net
Towards a Formal Theory of Representational Compositionality
Eric Elmoznino
Thomas Jiralerspong
Yoshua Bengio
Guillaume Lajoie
Compositionality is believed to be fundamental to intelligence. In humans, it underlies the structure of thought and language. In AI, it ena… (voir plus)bles a powerful form of out-of-distribution generalization, in which a model systematically adapts to novel combinations of known concepts. However, while we have strong intuitions about what compositionality is, we lack satisfying formal definitions for it. Here, we propose such a definition called representational compositionality that is conceptually simple, quantitative, and grounded in algorithmic information theory. Intuitively, representational compositionality states that a compositional representation is both expressive and describable as a simple function of parts. We validate our definition on both real and synthetic data, and show how it unifies disparate intuitions from across the literature in both AI and cognitive science. We hope that our definition can inspire the design of novel, theoretically-driven models that better capture the mechanisms of compositional thought. We make our code available at https://github.com/EricElmoznino/complexity_compositionality.
2025-05-01
ICML.cc/2025/Conference (poster)
openreview.net
openreview.net
TypyBench: Evaluating LLM Type Inference for Untyped Python Repositories
Honghua Dong
Jiacheng Yang
Xun Deng
Yuhe Jiang
Gennady Pekhimenko
Fan Long
Xujie Si
2025-05-01
ICML.cc/2025/Conference (poster)
openreview.net
openreview.net
UI-Vision: A Desktop-centric GUI Benchmark for Visual Perception and Interaction
Shravan Nayak
Xiangru Jian
Kevin Qinghong Lin
Juan A. Rodriguez
Montek Kalsi
Rabiul Awal
Nicolas Chapados
M. Tamer Özsu
Aishwarya Agrawal
David Vazquez
Chris Pal
Perouz Taslakian
Spandana Gella
Sai Rajeswar
Human Annotator
2025-05-01
ICML.cc/2025/Conference (poster)
doi.org
openreview.net
Uncovering a Universal Abstract Algorithm for Modular Addition in Neural Networks
Gavin McCracken
Gabriela Moisescu-Pareja
Vincent Létourneau
Doina Precup
Jonathan Love
We propose a testable universality hypothesis, asserting that seemingly disparate neural network solutions observed in the simple task of mo… (voir plus)dular addition are unified under a common abstract algorithm. While prior work interpreted variations in neuron-level representations as evidence for distinct algorithms, we demonstrate - through multi-level analyses spanning neurons, neuron clusters, and entire networks - that multilayer perceptrons and transformers universally implement the abstract algorithm we call the approximate Chinese Remainder Theorem. Crucially, we introduce approximate cosets and show that neurons activate exclusively on them. Furthermore, our theory works for deep neural networks (DNNs). It predicts that universally learned solutions in DNNs with trainable embeddings or more than one hidden layer require only O(log n) features, a result we empirically confirm. This work thus provides the first theory-backed interpretation of multilayer networks solving modular addition. It advances generalizable interpretability and opens a testable universality hypothesis for group multiplication beyond modular addition.
2025-05-01
arXiv (publié)
doi.org
arxiv.org
VinePPO: Refining Credit Assignment in RL Training of LLMs
Amirhossein Kazemnejad
Milad Aghajohari
Eva Portelance
Alessandro Sordoni
Siva Reddy
Aaron Courville
Nicolas Le Roux
Large language models (LLMs) are increasingly applied to complex reasoning tasks that require executing several complex steps before receivi… (voir plus)ng any reward. Properly assigning credit to these steps is essential for enhancing model performance. Proximal Policy Optimization (PPO), a common reinforcement learning (RL) algorithm used for LLM finetuning, employs value networks to tackle credit assignment. However, recent approaches achieve strong results without it, raising questions about the efficacy of value networks in practice. In this work, we systematically evaluate the efficacy of value networks and reveal their significant shortcomings in reasoning-heavy LLM tasks, showing that they often produce poor estimate of expected return and barely outperform a random baseline when comparing alternative steps. This motivates our key question: Can improved credit assignment enhance RL training for LLMs? To address this, we propose VinePPO, a straightforward approach that leverages the flexibility of language environments to compute unbiased Monte Carlo-based estimates. Our method consistently outperforms PPO and other baselines across MATH and GSM8K datasets in less wall-clock time (up to 3.0x). Crucially, it achieves higher test accuracy for a given training accuracy, capturing more generalization signal per sample. These results emphasize the importance of accurate credit assignment in RL training of LLM.
2025-05-01
ICML.cc/2025/Conference (poster)
openreview.net
openreview.net
Virtual Cells: Predict, Explain, Discover
Emmanuel Noutahi
Jason Hartford
Prudencio Tossou
Shawn Whitfield
Ali Denton
Cas Wognum
Kristina Ulicna
Michael Craig
Jonathan Hsu
Michael Cuccarese
Emmanuel Bengio
Dominique Beaini
Christopher Gibson
Daniel Cohen
Berton Earnshaw
2025-05-01
arXiv (publié)
doi.org
arxiv.org
When to retrain a machine learning model
Florence Regol
Leo Schwinn
Kyle Sprague
Mark Coates
Thomas Markovich
A significant challenge in maintaining real-world machine learning models is responding to the continuous and unpredictable evolution of dat… (voir plus)a. Most practitioners are faced with the difficult question: when should I retrain or update my machine learning model? This seemingly straightforward problem is particularly challenging for three reasons: 1) decisions must be made based on very limited information - we usually have access to only a few examples, 2) the nature, extent, and impact of the distribution shift are unknown, and 3) it involves specifying a cost ratio between retraining and poor performance, which can be hard to characterize. Existing works address certain aspects of this problem, but none offer a comprehensive solution. Distribution shift detection falls short as it cannot account for the cost trade-off; the scarcity of the data, paired with its unusual structure, makes it a poor fit for existing offline reinforcement learning methods, and the online learning formulation overlooks key practical considerations. To address this, we present a principled formulation of the retraining problem and propose an uncertainty-based method that makes decisions by continually forecasting the evolution of model performance evaluated with a bounded metric. Our experiments, addressing classification tasks, show that the method consistently outperforms existing baselines on 7 datasets. We thoroughly assess its robustness to varying cost trade-off values and mis-specified cost trade-offs.
2025-05-01
ICML.cc/2025/Conference (poster)
openreview.net
openreview.net
Caffeine induces age-dependent increases in brain complexity and criticality during sleep
Philipp Thölke
Maxine Arcand-Lavigne
Tarek Lajnef
Sonia Frenette
Julie Carrier
Karim Jerbi
2025-04-30
Communications Biology (publié)
doi.org
ImmunoStruct: a multimodal neural network framework for immunogenicity prediction from peptide-MHC sequence, structure, and biochemical properties
Kevin Bijan Givechian
João Felipe Rocha
Edward Yang
Chen Liu
Kerrie Greene
Rex Ying
Etienne Caron
Akiko Iwasaki
Smita Krishnaswamy
2025-04-30
bioRxiv (prépublication)
doi.org
JPerfEvo: A Tool for Tracking Method-Level Performance Changes in Java Projects
Kaveh Shahedi
Maxime Lamothe
Foutse Khomh
Heng Li
Performance regressions and improvements are common phenomena in software development, occurring periodically as software evolves and mature… (voir plus)s. When developers introduce new changes to a program’s codebase, unforeseen performance variations may arise. Identifying these changes at the method level, however, can be challenging due to the complexity and scale of modern codebases. In this work, we present JPerfEvo, a tool designed to automate the evaluation of the method-level performance impact of each code commit (i.e., the performance variations between the two versions before and after a commit). Leveraging the Java Microbenchmark Harness (JMH) module for benchmarking the modified methods, JPerfEvo instruments their execution and applies robust statistical evaluations to detect performance changes. The tool can classify these changes as performance improvements, regressions, or neutral (i.e., no change), with the change magnitude. We evaluated JPerfEvo on three popular and mature open-source Java projects, demonstrating its effectiveness in identifying performance changes throughout their development histories.
2025-04-28
IEEE Working Conference on Mining Software Repositories (publié)
doi.org
Proceedings of 1st Workshop on Advancing Artificial Intelligence through Theory of Mind
Mouad Abrini
Omri Abend
Dina M. Acklin
Henny Admoni
Gregor Aichinger
Nitay Alon
Zahra Ashktorab
Ashish Atreja
Moises Auron
Alexander Aufreiter
Raghav Awasthi
Soumya Banerjee
Joseph Barnby
Rhea Basappa
Severin Bergsmann
Djallel Bouneffouf
Patrick Callaghan
Marc Cavazza
Thierry Chaminade
Sonia Chernova … (voir 88 de plus)
Mohamed Chetouan
Moumita Choudhury
Axel Cleeremans
J. Cywinski
Fabio Cuzzolin
Hokin Deng
N'yoma Diamond
C. D. Pasquasio
Guillaume Dumas
Max J. van Duijn
Mahapatra Dwarikanath
Qingying Gao
Ashok Goel
Rebecca R. Goldstein
Matthew C. Gombolay
Gabriel Enrique Gonzalez
Amar Halilovic
Tobias Halmdienst
Mahimul Islam
Julian Jara-Ettinger
Natalie Kastel
Renana Keydar
Ashish K. Khanna
Mahdi Khoramshahi
Jihyun Kim
Mihyeon Kim
Youngbin Kim
Senka Krivic
Nikita Krasnytskyi
Arun Kumar
Junehyoung Kwon
EunJu Lee
Shane Lee
Peter R. Lewis 0001
Xue Li
Yijiang Li
Michal Lewandowski
Nathan Lloyd
Matthew B. Luebbers
Dezhi Luo
Haiyun Lyu
Dwarikanath Mahapatra
Kamal Maheshwari
Mallika Mainali
P. Mathur
Patrick Mederitsch
Shuwa Miura
Manuel Preston de Miranda
Reuth Mirsky
Shreya Mishra
Nina M. Moorman
Katelyn Morrison
John Muchovej
Bernhard Nessler
Felix Nessler
Hieu Minh Jord Nguyen
Abby Ortego
F. Papay
Antoine Pasquali
Hamed Rahimi
C. Raghu
Amanda L. Royka
Stefan Sarkadi
Jaelle Scheuerman
Simon Schmid
Paul Schrater
Anik Sen
Zahra Sheikhbahaee
Ke Shi
Reid G. Simmons
Nishant Singh
Mason O. Smith
Ramira van der Meulen
Anthia Solaki
Haoran Sun
Viktor Szolga
Matthew E. Taylor
Travis Taylor
Sanne van Waveren
Juan David Vargas
R. Verbrugge
Eitan Wagner
Justin D. Weisz
Ximing Wen
William Yeoh
Wenlong Zhang
Michelle Zhao
Shlomo Zilberstein
2025-04-28
ArXiv (prépublication)
arxiv.org
arxiv.org