Publications

Action abstractions for amortized sampling

Oussama Boussif

Lena Nehale Ezzine

Joseph D Viviano

Michał Koziarski

Moksh J. Jain

Nikolay Malkin

Emmanuel Bengio

Rim Assouel

2025-01-22

ICLR.cc/2025/Conference (poster)

AdaFisher: Adaptive Second Order Optimization via Fisher Information

Damien MARTINS GOMES

Yanlei Zhang

Eugene Belilovsky

Guy Wolf

Mahdi S. Hosseini

First-order optimization methods are currently the mainstream in training deep neural networks (DNNs). Optimizers like Adam incorporate limi… (see more)ted curvature information by employing the diagonal matrix preconditioning of the stochastic gradient during the training. Despite their widespread, second-order optimization algorithms exhibit superior convergence properties compared to their first-order counterparts e.g. Adam and SGD. However, their practicality in training DNNs are still limited due to increased per-iteration computations and suboptimal accuracy compared to the first order methods. We present AdaFisher--an adaptive second-order optimizer that leverages a block-diagonal approximation to the Fisher information matrix for adaptive gradient preconditioning. AdaFisher aims to bridge the gap between enhanced convergence capabilities and computational efficiency in second-order optimization framework for training DNNs. Despite the slow pace of second-order optimizers, we showcase that AdaFisher can be reliably adopted for image classification, language modelling and stand out for its stability and robustness in hyperparameter tuning. We demonstrate that AdaFisher outperforms the SOTA optimizers in terms of both accuracy and convergence speed. Code available from \href{https://github.com/AtlasAnalyticsLab/AdaFisher}{https://github.com/AtlasAnalyticsLab/AdaFisher}

2025-01-22

ICLR.cc/2025/Conference (poster)

Adaptive teachers for amortized samplers

Minsu Kim

Sanghyeok Choi

Taeyoung Yun

Emmanuel Bengio

Leo Feng

Jarrid Rector-Brooks

Sungsoo Ahn

Jinkyoo Park

Nikolay Malkin

Amortized inference is the task of training a parametric model, such as a neural network, to approximate a distribution with a given unnorma… (see more)lized density where exact sampling is intractable. When sampling is implemented as a sequential decision-making process, reinforcement learning (RL) methods, such as generative flow networks, can be used to train the sampling policy. Off-policy RL training facilitates the discovery of diverse, high-reward candidates, but existing methods still face challenges in efficient exploration. We propose to use an adaptive training distribution (the Teacher) to guide the training of the primary amortized sampler (the Student) by prioritizing high-loss regions. The Teacher, an auxiliary behavior model, is trained to sample high-error regions of the Student and can generalize across unexplored modes, thereby enhancing mode coverage by providing an efficient training curriculum. We validate the effectiveness of this approach in a synthetic environment designed to present an exploration challenge, two diffusion-based sampling tasks, and four biochemical discovery tasks demonstrating its ability to improve sample efficiency and mode coverage.

2025-01-22

ICLR.cc/2025/Conference (poster)

Advantage Alignment Algorithms

Juan Agustin Duque

Milad Aghajohari

Tim Cooijmans

razvan ciuca

Tianyu Zhang

Gauthier Gidel

Aaron Courville

2025-01-22

ICLR.cc/2025/Conference (oral)

AFlow: Automating Agentic Workflow Generation

Jiayi Zhang

Jinyu Xiang

Zhaoyang Yu

Fengwei Teng

Xiong-Hui Chen

Jiaqi Chen

Mingchen Zhuge

Xin Cheng

Sirui Hong

Jinlin Wang

Bingnan Zheng

Bang Liu

Yuyu Luo

Chenglin Wu

Large language models (LLMs) have demonstrated remarkable potential in solving complex tasks across diverse domains, typically by employing … (see more)agentic workflows that follow detailed instructions and operational sequences. However, constructing these workflows requires significant human effort, limiting scalability and generalizability. Recent research has sought to automate the generation and optimization of these workflows, but existing methods still rely on initial manual setup and fall short of achieving fully automated and effective workflow generation. To address this challenge, we reformulate workflow optimization as a search problem over code-represented workflows, where LLM-invoking nodes are connected by edges. We introduce AFLOW, an automated framework that efficiently explores this space using Monte Carlo Tree Search, iteratively refining workflows through code modification, tree-structured experience, and execution feedback. Empirical evaluations across six benchmark datasets demonstrate AFLOW's efficacy, yielding a 5.7% average improvement over state-of-the-art baselines. Furthermore, AFLOW enables smaller models to outperform GPT-4o on specific tasks at 4.55% of its inference cost in dollars. The code is available at https://github.com/geekan/MetaGPT.

2025-01-22

ICLR.cc/2025/Conference (oral)

Ant Colony Sampling with GFlowNets for Combinatorial Optimization

Minsu Kim

Sanghyeok Choi

Jiwoo Son

Hyeonah Kim

Jinkyoo Park

2025-01-22

aistats.org/AISTATS/2025/Conference (poster)

AssembleFlow: Rigid Flow Matching with Inertial Frames for Molecular Assembly

Hongyu Guo

Shengchao Liu

Molecular assembly, where a cluster of rigid molecules aggregated into strongly correlated forms, is fundamental to determining the properti… (see more)es of materials. However, traditional numerical methods for simulating this process are computationally expensive, and existing generative models on material generation overlook the rigidity inherent in molecular structures, leading to unwanted distortions and invalid internal structures in molecules. To address this, we introduce AssembleFlow. AssembleFlow leverages inertial frames to establish reference coordinate systems at the molecular level for tracking the orientation and motion of molecules within the cluster. It further decomposes molecular

2025-01-22

ICLR.cc/2025/Conference (poster)

Asynchronous RLHF: Faster and More Efficient Off-Policy RL for Language Models

Michael Noukhovitch

Shengyi Huang

Sophie Xhonneux

Arian Hosseini

Rishabh Agarwal

Aaron Courville

The dominant paradigm for RLHF is online and on-policy RL: synchronously generating from the large language model (LLM) policy, labelling wi… (see more)th a reward model, and learning using feedback on the LLM's own outputs. While performant, this paradigm is computationally inefficient. Inspired by classical deep RL literature, we propose separating generation and learning in RLHF. This enables asynchronous generation of new samples while simultaneously training on old samples, leading to faster training and more compute-optimal scaling. However, asynchronous training relies on an underexplored regime, online but off-policy RLHF: learning on samples from previous iterations of our model. To understand the challenges in this regime, we investigate a fundamental question: how much off-policyness can we tolerate for asynchronous training to speed up learning but maintain performance? Among several RLHF algorithms we tested, we find that online DPO is most robust to off-policy data, and robustness increases with the scale of the policy model. We study further compute optimizations for asynchronous RLHF but find that they come at a performance cost, giving rise to a trade-off. Finally, we verify the scalability of asynchronous RLHF by training LLaMA 3.1 8B on an instruction-following task 40% faster than a synchronous run while matching final performance.

2025-01-22

ICLR.cc/2025/Conference (poster)

Beyond FVD: An Enhanced Evaluation Metrics for Video Generation Distribution Quality

Ge Ya Luo

Gian Mario Favero

Zhi Hao Luo

Alexia Jolicoeur-Martineau

Chris Pal

The Fréchet Video Distance (FVD) is a widely adopted metric for evaluating video generation distribution quality. However, its effectivenes… (see more)s relies on critical assumptions. Our analysis reveals three significant limitations: (1) the non-Gaussianity of the Inflated 3D Convnet (I3D) feature space; (2) the insensitivity of I3D features to temporal distortions; (3) the impractical sample sizes required for reliable estimation. These findings undermine FVD's reliability and show that FVD falls short as a standalone metric for video generation evaluation. After extensive analysis of a wide range of metrics and backbone architectures, we propose JEDi, the JEPA Embedding Distance, based on features derived from a Joint Embedding Predictive Architecture, measured using Maximum Mean Discrepancy with polynomial kernel. Our experiments on multiple open-source datasets show clear evidence that it is a superior alternative to the widely used FVD metric, requiring only 16% of the samples to reach its steady value, while increasing alignment with human evaluation by 34%, on average. Project page: https://oooolga.github.io/JEDi.github.io/.

2025-01-22

ICLR.cc/2025/Conference (poster)

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

Juan A. Rodriguez

Xiangru Jian

Siba Smarak Panigrahi

Tianyu Zhang

Aarash Feizi

Abhay Puri

Akshay Kalkunte Suresh

François Savard

Ahmed Masry

Shravan Nayak

Rabiul Awal

Mahsa Massoud

Amirhossein Abaskohi

Zichao Li

Suyuchen Wang

Pierre-Andre Noel

Mats Leon Richter

Saverio Vadacchino

Shubham Agarwal

Sanket Biswas … (see 23 more)

Sara Shanian

Ying Zhang

Noah Bolger

Kurt MacDonald

Simon Fauvel

Sathwik Tejaswi Madhusudhan

Srinivas Sunkara

Joao Monteiro

Krishnamurthy Dj Dvijotham

Torsten Scholak

Nicolas Chapados

Sepideh Kharaghani

Sean Hughes

M. Özsu

Issam Hadj Laradji

Spandana Gella

Perouz Taslakian

David Vazquez

Sai Rajeswar

2025-01-22

ICLR.cc/2025/Conference (poster)

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

Juan A. Rodriguez

Xiangru Jian

Siba Smarak Panigrahi

Tianyu Zhang

Aarash Feizi

Abhay Puri

Akshay Kalkunte Suresh

François Savard

Ahmed Masry

Shravan Nayak

Rabiul Awal

Mahsa Massoud

Amirhossein Abaskohi

Zichao Li

Suyuchen Wang

Pierre-Andre Noel

Mats Leon Richter

Saverio Vadacchino

Shubham Agarwal

Sanket Biswas … (see 19 more)

Sara Shanian

Ying Zhang

Sathwik Tejaswi Madhusudhan

Joao Monteiro

Krishnamurthy Dj Dvijotham

Torsten Scholak

Nicolas Chapados

Sepideh Kharaghani

Sean Hughes

M. Özsu

Issam Hadj Laradji

Spandana Gella

Perouz Taslakian

David Vazquez

Sai Rajeswar

Multimodal AI has the potential to significantly enhance document-understanding tasks, such as processing receipts, understanding workflows,… (see more) extracting data from documents, and summarizing reports. Code generation tasks that require long-structured outputs can also be enhanced by multimodality. Despite this, their use in commercial applications is often limited due to limited access to relevant training data and restrictive licensing, which hinders open access. To address these limitations, we introduce BigDocs-7.5M, a high-quality, open-access dataset comprising 7.5 million multimodal documents across 30 tasks. We use an efficient data curation process to ensure that our data is high quality and license-permissive. Our process emphasizes accountability, responsibility, and transparency through filtering rules, traceable metadata, and careful content analysis. Additionally, we introduce BigDocs-Bench,, a benchmark suite with 10 novel tasks where we carefully create datasets that reflect real-world use cases involving reasoning over Graphical User Interfaces (GUI) and code generation from images. Our experiments show that training with BigDocs-Bench, improves average performance up to 25.8% over closed-source GPT-4o in document reasoning and structured output tasks such as Screenshot2HTML or Image2Latex generation. Finally, human evaluations revealed that participants preferred the outputs from models trained with BigDocs over those from GPT-4o. This suggests that BigDocs can help both academics and the open-source community utilize and improve AI tools to enhance multimodal capabilities and document reasoning.

2025-01-22

ICLR.cc/2025/Conference (poster)