1.1k citations · 4.4k across the 113 of their papers we have counts for
35 papers · 1 filter
Reinforcement Learning for Code Optimization
Pierre Chambon, Kunhao Zheng, Juliette Decugis +2
RL for code correctness is now established: have the model generate a program, run it against hidden test cases, and reward solutions that pass. Extending this to code optimization…
Sparse Delta Memory: Scaling the State of Linear RNNs through Sparsity
Loïc Cabannes, Pierre-Emmanuel Mazaré, Gergely Szilvasy +6
Linear attention models allow a fixed state size and a fixed amount of compute per token. However, due to their limited state size, linear attention models fall behind in long-cont…
DecompRL: Solving Harder Problems by Learning Modular Code Generation
Juliette Decugis, Fabian Gloeckle, Francis Bach +2
How can Large Language Models (LLMs) solve problems they currently cannot? Repeated sampling scales test-time compute but GPU cost grows linearly with attempts, while reinforcement…
Don't Let Gains FADE: Breaking Down Policy Gradient Weights in RL
Juliette Decugis, Sean O'Brien, Francis Bach +2
Reinforcement learning post-training dramatically improves LLM reasoning, but suffers from training instability and diversity collapse. Advantage functions offer an appealing fix:…
Extrapolative Weight Averaging Reveals Correctness-Efficiency Frontiers in Code RL
Kunhao Zheng, Pierre Chambon, Juliette Decugis +4
Linear interpolation between fine-tuned checkpoints has been shown to trace the Pareto front between competing objectives, but whether extrapolative weight averaging can extend suc…
A Deep Dive into Scaling RL for Code Generation with Synthetic Data and Curricula
Cansu Sancaktar, David Zhang, Gabriel Synnaeve +1
Reinforcement learning (RL) has emerged as a powerful paradigm for improving large language models beyond supervised fine-tuning, yet sustaining performance gains at scale remains…
Towards a Neural Debugger for Python
Maximilian Beck, Jonas Gehring, Jannik Kossen +1
Training large language models (LLMs) on Python execution traces grounds them in code execution and enables the line-by-line execution prediction of whole Python programs, effectiv…
Set Block Decoding is a Language Model Inference Accelerator
Itai Gat, Heli Ben-Hamu, Marton Havasi +6
Autoregressive next token prediction language models offer powerful capabilities but face significant challenges in practical deployment due to the high computational and memory co…
Short window attention enables long-term memorization
Loïc Cabannes, Maximilian Beck, Gergely Szilvasy +6
Recent works show that hybrid architectures combining local sliding window attention layers and global attention layers outperform either of these architectures taken separately. H…
Stochastic activations
Maria Lomeli, Matthijs Douze, Gergely Szilvasy +7
We introduce stochastic activations. This novel strategy randomly selects between several non-linear functions in the feed-forward layer of a large language model. In particular, w…
Soft Policy Optimization: Online Off-Policy RL for Sequence Models
Taco Cohen, David W. Zhang, Kunhao Zheng +3
RL-based post-training of language models is almost exclusively done using on-policy methods such as PPO. These methods cannot learn from arbitrary sequences such as those produced…
Optimizing Language Models for Inference Time Objectives using Reinforcement Learning
Yunhao Tang, Kunhao Zheng, Gabriel Synnaeve +1
In this work, we investigate the merits of explicitly optimizing for inference time algorithmic performance during model training. We show how optimizing for inference time perform…
What I cannot execute, I do not understand: Training and Evaluating LLMs on Program Execution Traces
Jordi Armengol-Estapé, Quentin Carbonneaux, Tianjun Zhang +8
Code generation and understanding are critical capabilities for large language models (LLMs). Thus, most LLMs are pretrained and fine-tuned on code data. However, these datasets ty…
Don't Transform the Code, Code the Transforms: Towards Precise Code Rewriting using LLMs
Chris Cummins, Volker Seeker, Jordi Armengol-Estapé +3
Tools for rewriting, refactoring and optimizing code should be fast and correct. Large language models (LLMs), by their nature, possess neither of these qualities. Yet, there remai…
Discrete Flow Matching
Itai Gat, Tal Remez, Neta Shaul +5
Despite Flow Matching and diffusion models having emerged as powerful generative paradigms for continuous variables such as images and videos, their application to high-dimensional…
Priority Sampling of Large Language Models for Compilers
Dejan Grubisic, Chris Cummins, Volker Seeker +1
Large language models show great potential in generating and optimizing code. Widely used sampling methods such as Nucleus Sampling increase the diversity of generation but often p…
A Data Source for Reasoning Embodied Agents
Jack Lanchantin, Sainbayar Sukhbaatar, Gabriel Synnaeve +3
Recent progress in using machine learning models for reasoning tasks has been driven by novel model architectures, large-scale pre-training protocols, and dedicated reasoning datas…
Leveraging Demonstrations with Latent Space Priors
Jonas Gehring, Deepak Gopinath, Jungdam Won +3
Demonstrations provide insight into relevant state or action space regions, bearing great potential to boost the efficiency and practicality of reinforcement learning agents. In th…
Star Temporal Classification: Sequence Classification with Partially Labeled Data
Vineel Pratap, Awni Hannun, Gabriel Synnaeve +1
We develop an algorithm which can learn from partially labeled and unsegmented sequential data. Most sequential loss functions, such as Connectionist Temporal Classification (CTC),…
Flashlight: Enabling Innovation in Tools for Machine Learning
Jacob Kahn, Vineel Pratap, Tatiana Likhomanenko +11
As the computational requirements for machine learning systems and the size and complexity of machine learning frameworks increases, essential framework innovation has become chall…
Hierarchical Skills for Efficient Exploration
Jonas Gehring, Gabriel Synnaeve, Andreas Krause +1
In reinforcement learning, pre-trained low-level skills have the potential to greatly facilitate exploration. However, prior knowledge of the downstream task is required to strike…
CAPE: Encoding Relative Positions with Continuous Augmented Positional Embeddings
Tatiana Likhomanenko, Qiantong Xu, Gabriel Synnaeve +2
Without positional information, attention-based Transformer neural networks are permutation-invariant. Absolute or relative positional embeddings are the most popular ways to feed…
Gradient Matching for Domain Generalization
Yuge Shi, Jeffrey Seely, Philip H. S. Torr +4
Machine learning systems typically assume that the distributions of training and test sets match closely. However, a critical requirement of such systems in the real world is their…
Self-training and Pre-training are Complementary for Speech Recognition
Qiantong Xu, Alexei Baevski, Tatiana Likhomanenko +5
Self-training and unsupervised pre-training have emerged as effective approaches to improve speech recognition systems using unlabeled data. However, it is not clear whether they l…