Yoshua Bengio

FALCON: Few-step Accurate Likelihoods for Continuous Flows

Efficient Regression-based Training of Normalizing Flows for Boltzmann Generators

Visual symbolic mechanisms: Emergent symbol processing in Vision Language Models

Latent Veracity Inference for Identifying Errors in Stepwise Reasoning

A Comedy of Estimators: On KL Regularization in RL Training of LLMs

Active Attacks: Red-teaming LLMs via Adaptive Environments

In-Context Reinforcement Learning through Bayesian Fusion of Context and Value Prior

Recursive Self-Aggregation Unlocks Deep Thinking in Large Language Models

Discrete Feynman-Kac Correctors

Structure-Aligned Protein Language Model

SCOPE: Selective Cross-modal Orchestration of Visual Perception Experts

Torsional-GFN: a conditional conformation generator for small molecules

Learning What Matters: Steering Diffusion via Spectrally Anisotropic Forward Noise

Bayesian Symbolic Regression with Entropic Reinforcement Learning

Self-Evolving Curriculum for LLM Reasoning

Surrogate-Based Quantification of Policy Uncertainty in Generative Flow Networks

Fast Monte Carlo Tree Diffusion: 100× Speedup via Parallel and Sparse Planning

Adaptive Inference-Time Scaling via Cyclic Diffusion Search

Bringing SAM to new heights: leveraging elevation data for tree crown segmentation from drone imagery

Trajectory Balance with Asynchrony: Decoupling Exploration and Learning for Fast, Scalable LLM Post-Training

Structure Language Models for Protein Conformation Generation

HarmAug: Effective Data Augmentation for Knowledge Distillation of Safety Guard Models

Learning Diverse Attacks on Large Language Models for Robust Red-Teaming and Safety Tuning

Action abstractions for amortized sampling

AlignVLM: Bridging Vision and Language Latent Spaces for Multimodal Document Understanding

Towards a formal theory of compositionality

VCR: A Task for Pixel-Level Complex Reasoning in Vision Language Models via Restoring Occluded Text

Outsourced Diffusion Sampling: Efficient Posterior Inference in Latent Spaces of Generative Models

AI for Global Climate Cooperation: Modeling Global Climate Negotiations, Agreements, and Long-Term Cooperation in RICE-N

AssembleFlow: Rigid Flow Matching with Inertial Frames for Molecular Assembly

Towards Improving Exploration through Sibling Augmented GFlowNets

Adaptive teachers for amortized samplers

MAP: Low-compute Model Merging with Amortized Pareto Fronts via Quadratic Approximation

RL, but don't do anything I wouldn't do

Meta Flow Matching: Integrating Vector Fields on the Wasserstein Manifold

Towards a Formal Theory of Representational Compositionality

Monte Carlo Tree Diffusion for System 2 Planning

Efficient Diversity-Preserving Diffusion Alignment via Gradient-Informed GFlowNets

Low Compute Unlearning via Sparse Representations

BigDocs: An Open Dataset for Training Multimodal Models on Document and Code Tasks

GEOMETRIC SIGNATURES OF COMPOSITIONALITY ACROSS A LANGUAGE MODEL’S LIFETIME

Can a Bayesian oracle prevent harm from an agent?

Identifying and Addressing Delusions for Target-Directed Decision Making

Hierarchical Multiscale Diffuser for Extendable Long-Horizon Planning

Were RNNs All We Needed?

On the Transfer of Object-Centric Representation Learning

Rejecting Hallucinated State Targets during Planning

Mitigating Shortcut Learning with Diffusion Counterfactuals and Diverse Ensembles

AI-Assisted Generation of Difficult Math Questions