Yoshua Bengio

Scaling in the Service of Reasoning & Model-Based ML

April 4, 2023

Yoshua Bengio

Edward J. Hu

A collaboration between Mila and Relation Therapeutics to discover novel synergistic combinations of drugs in vitro

March 23, 2022

Paul Bertin

Jake P. Taylor-King

Yoshua Bengio

March 15, 2022

Generative Flow Networks

Yoshua Bengio

Publications

Sliding Window Recurrences for Sequence Models

Dragos Secrieru

Garyk Brixi

Taiji Suzuki

Michael Poli

Stefano Massaroli

Multi-hybrid architectures are poised to take over language modeling due to better quality and performance. We introduce a hierarchical deco… (see more)mposition framework for linear recurrences that allows us to develop algorithms aligned with GPU memory hierarchies, yielding Sliding Window Recurrences. We focus specifically on truncating recurrences to hardware-aligned windows which are naturally jagged, limiting costly inter-warp communication. Using SWR, we develop Phalanx layers that serve as drop-in replacements for windowed attention or linear recurrences. In 1B parameter multi-hybrid models, Phalanx achieves over 10-40% speedup across 4K to 32K context length over optimized Transformers while matching perplexity.

2025-11-30

arXiv (published)

Leveraging a Fully Differentiable Integrated Assessment Model for RL and Inference

Koen Ponse

Kai-Hendrik Cohrs

Phillip Wozny

Andrew Robert Williams

Tianyu Zhang

Erman Acar

Aske Plaat

Thomas M. Moerland

Pierre Gentine

Gustau Camps-Valls

2025-11-20

EurIPS.cc/2025/Workshop/DiffSys (published)

openreview.net

A HOT Dataset: 150,000 Buildings for HVAC Operations Transfer Research

Anaïs Berkes

David Rolnick

Donna Vakalis

About 12% of global energy consumption is attributable to heating, ventilation, and air conditioning (HVAC) systems in buildings [11]. Machi… (see more)ne learning-based intelligent HVAC control offers significant energy efficiency potential, but progress is constrained by limited data for training and evaluating performance across different kinds of buildings. Existing datasets primarily target energy prediction rather than control applications, forcing studies to rely on limited building sets or single-variable perturbations that fail to capture real-world complexity. We present HOT (HVAC Operations Transfer), the first large-scale open-source dataset purpose-built for research into transfer learning in building control. HOT contains 159,744 unique building-weather combinations with systematic variations across envelope properties, occupancy patterns, and climate conditions spanning all 19 ASHRAE climate zones across 76 global locations. We formalise a comprehensive similarity-based framework with quantitative metrics for assessing transfer feasibility between source and target buildings across multiple context dimensions. Our key contributions: (1) a large-scale, open dataset and tooling enabling systematic, multi-variable transfer studies across 19 climate zones; (2) a quantitative similarity framework spanning geometry, thermal, climate, and function; and (3) zero-shot climate transfer experiments showing why realistic context variation matters for HVAC control.

2025-11-10

Proceedings of the 12th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation (published)

A HOT Dataset: 150,000 Buildings for HVAC Operations Transfer Research

Anaïs Berkes

David Rolnick

Donna Vakalis

2025-11-10

Proceedings of the 12th ACM International Conference on Systems for Energy-Efficient Buildings, Cities, and Transportation (published)

Scaling Latent Reasoning via Looped Language Models

Ruiming Zhu

Zixuan Wang

Kai Hua

Tianyu Zhang

Ziniu Li

Haoran Que

Boyi Wei

Zixin Wen

Fan Yin

He Xing

Li Li

Jiajun Shi

Kaijing Ma

Shanda Li

Taylor Kergan

Andrew C. Smith

Xin Qu

Mude Hui

Bohong Wu

Qiyang Min … (see 13 more)

Hongzhi Huang

Xun Zhou

Wei Ye

Jiaheng Liu

Jian Yang 0030

Yunfeng Shi

Chenghua Lin

Enduo Zhao

Tianle Cai

Ge Zhang

Wenhao Huang

Jason K. Eshraghian

Modern LLMs are trained to"think"primarily via explicit text generation, such as chain-of-thought (CoT), which defers reasoning to post-trai… (see more)ning and under-leverages pre-training data. We present and open-source Ouro, named after the recursive Ouroboros, a family of pre-trained Looped Language Models (LoopLM) that instead build reasoning into the pre-training phase through (i) iterative computation in latent space, (ii) an entropy-regularized objective for learned depth allocation, and (iii) scaling to 7.7T tokens. Ouro 1.4B and 2.6B models enjoy superior performance that match the results of up to 12B SOTA LLMs across a wide range of benchmarks. Through controlled experiments, we show this advantage stems not from increased knowledge capacity, but from superior knowledge manipulation capabilities. We also show that LoopLM yields reasoning traces more aligned with final outputs than explicit CoT. We hope our results show the potential of LoopLM as a novel scaling direction in the reasoning era. Our model is available here: http://ouro-llm.github.io.

2025-10-28

ArXiv (preprint)

Deep-learning-based virtual screening of antibacterial compounds

Gabriele Scalia

Steven T. Rutherford

Ziqing Lu

Kerry R. Buchholz

Nicholas Skelton

Kangway Chuang

Nathaniel Diamant

Jan-Christian Hütter

Jerome-Maxim Luescher

Anh Miu

Jeff Blaney

Leo Gendelev

Elizabeth Skippington

Greg Zynda

Nia Dickson

Michał Koziarski

Aviv Regev

Man-Wah Tan

Tommaso Biancalani

2025-10-23

Nature Biotechnology (unknown)

Surrogate-based quantification of policy uncertainty in generative flow networks

Ram'on Nartallo-Kaluarachchi

Robert Manson-Sawko

Shashanka Ubaru

Dongsung Huh

Malgorzata J. Zimo'n

Lior Horesh

2025-10-23

ArXiv (preprint)

Monte Carlo Tree Diffusion for System 2 Planning

Jaesik Yoon

Hyeonseo Cho

Doojin Baek

Sungjin Ahn

Diffusion models have recently emerged as a powerful tool for planning. However, unlike Monte Carlo Tree Search (MCTS)-whose performance nat… (see more)urally improves with inference-time computation scaling-standard diffusion-based planners offer only limited avenues for the scalability. In this paper, we introduce Monte Carlo Tree Diffusion (MCTD), a novel framework that integrates the generative strength of diffusion models with the adaptive search capabilities of MCTS. Our method reconceptualizes denoising as a tree-structured process, allowing partially denoised plans to be iteratively evaluated, pruned, and refined. By selectively expanding promising trajectories while retaining the flexibility to revisit and improve suboptimal branches, MCTD achieves the benefits of MCTS such as controlling exploration-exploitation trade-offs within the diffusion framework. Empirical results on challenging long-horizon tasks show that MCTD outperforms diffusion baselines, yielding higher-quality solutions as inference-time computation increases.

2025-10-05

Proceedings of the 42nd International Conference on Machine Learning (published)

proceedings.mlr.press

Towards a Formal Theory of Representational Compositionality

2025-10-05

Proceedings of the 42nd International Conference on Machine Learning (published)

proceedings.mlr.press

Learning What Matters: Steering Diffusion via Spectrally Anisotropic Forward Noise

Luca Scimeca

Thomas Jiralerspong

Berton Earnshaw

Jason Hartford

2025-09-30

arXiv (published)

HVAC-SPICE: Value-Uncertainty In-Context RL with Thompson Sampling for Zero-Shot HVAC Control

Anaïs Berkes

Urban buildings consume 40\% of global energy, yet most rely on inefficient rule-based HVAC systems due to the impracticality of deploying a… (see more)dvanced controllers across diverse building stock. In-context reinforcement learning (ICRL) offers promise for rapid deployment without per-building training, but standard supervised learning objectives that maximise likelihood of training actions inherit behaviour-policy bias and provide weak exploration under the distribution shifts common when transferring across buildings and climates. We present SPICE (Sampling Policies In-Context with Ensemble uncertainty), a novel ICRL method specifically designed for zero-shot building control that addresses these fundamental limitations. SPICE introduces two key methodological innovations: (i) a propensity-corrected, return-aware training objective that prioritises high-advantage, high-uncertainty actions to enable improvement beyond suboptimal training demonstrations, and (ii) lightweight value ensembles with randomised priors that provide explicit uncertainty estimates for principled episode-level Thompson sampling. At deployment, SPICE samples one value head per episode and acts greedily, resulting in temporally coherent exploration without test-time gradients or building-specific models. We establish a comprehensive experimental protocol using the HOT dataset to evaluate SPICE across diverse building types and climate zones, focusing on the energy efficiency, occupant comfort, and zero-shot transfer capabilities that are critical for urban-scale deployment.

2025-09-29

NeurIPS.cc/2025/Workshop/UrbanAI (poster)

openreview.net

Recursive Self-Aggregation Unlocks Deep Thinking in Large Language Models

Johan Samir Obando Ceron

Brian R. Bartoldson

Bhavya Kailkhura

Guillaume Lajoie

Glen Berseth

Nikolay Malkin

Moksh J. Jain

Test-time scaling methods improve the capabilities of large language models (LLMs) by increasing the amount of compute used during inference… (see more) to make a prediction. Inference-time compute can be scaled in parallel by choosing among multiple independent solutions or sequentially through self-refinement. We propose Recursive Self-Aggregation (RSA), a test-time scaling method inspired by evolutionary methods that combines the benefits of both parallel and sequential scaling. Each step of RSA refines a population of candidate reasoning chains through aggregation of subsets to yield a population of improved solutions, which are then used as the candidate pool for the next iteration. RSA exploits the rich information embedded in the reasoning chains -- not just the final answers -- and enables bootstrapping from partially correct intermediate steps within different chains of thought. Empirically, RSA delivers substantial performance gains with increasing compute budgets across diverse tasks, model families and sizes. Notably, RSA enables Qwen3-4B-Instruct-2507 to achieve competitive performance with larger reasoning models, including DeepSeek-R1 and o3-mini (high), while outperforming purely parallel and sequential scaling strategies across AIME-25, HMMT-25, Reasoning Gym, LiveCodeBench-v6, and SuperGPQA. We further demonstrate that training the model to combine solutions via a novel aggregation-aware reinforcement learning approach yields significant performance gains. Code available at https://github.com/HyperPotatoNeo/RSA.

2025-09-29

ArXiv (preprint)