English
Related papers

Related papers: Entropy Ratio Clipping as a Soft Global Constraint…

200 papers

GRPO-style reinforcement learning (RL)-based LLM fine-tuning algorithms have recently gained popularity. Relying on heuristic trust-region approximations, however, they can lead to brittle optimization behavior, as global importance-ratio…

Machine Learning · Computer Science 2026-02-09 Doyeon Lee , Eunyi Lyou , Hyunsoo Cho , Sookyung Kim , Joonseok Lee , Jaemoo Choi

State entropy regularization has empirically shown better exploration and sample complexity in reinforcement learning (RL). However, its theoretical guarantees have not been studied. In this paper, we show that state entropy regularization…

Machine Learning · Computer Science 2025-12-02 Yonatan Ashlag , Uri Koren , Mirco Mutti , Esther Derman , Pierre-Luc Bacon , Shie Mannor

Large Language Models (LLMs) have made remarkable progress in enhancing step-by-step reasoning through reinforcement learning. However, the Group Relative Policy Optimization (GRPO) algorithm, which relies on sparse reward rules, often…

Artificial Intelligence · Computer Science 2025-07-30 Xingjian Zhang , Siwei Wen , Wenjun Wu , Lei Huang

In reinforcement learning (RL), it is often advantageous to consider additional constraints on the action space to ensure safety or action relevance. Existing work on such action-constrained RL faces challenges regarding effective policy…

Machine Learning · Computer Science 2025-12-01 Roland Stolz , Michael Eichelbeck , Matthias Althoff

Reinforcement learning (RL) has shown great promise in large language models (LLMs) post-training, which typically rely on token-level clipping to maintain stability during optimization. Despite the empirical success of GRPO-style methods,…

Computation and Language · Computer Science 2026-05-18 Jiakang Wang , Runze Liu , Qingpeng Cai , Lei Lin , Wenping Hu , Xiu Li , Fuzheng Zhang , Guorui Zhou , Kun Gai , Ling Pan

Reinforcement Learning (RL) has shown great potential in complex control tasks, particularly when combined with deep neural networks within the Actor-Critic (AC) framework. However, in practical applications, balancing exploration, learning…

Robotics · Computer Science 2026-02-25 Zhiwei Shang , Xinyi Yuan , Wenjun Huang , Yunduan Cui , Di Chen , Meixin Zhu

Entropy Regularisation is a widely adopted technique that enhances policy optimisation performance and stability. A notable form of entropy regularisation is augmenting the objective with an entropy term, thereby simultaneously optimising…

Machine Learning · Computer Science 2024-07-26 Jean Seong Bjorn Choe , Jong-Kook Kim

Reasoning ability has become a defining capability of Large Language Models (LLMs), with Reinforcement Learning with Verifiable Rewards (RLVR) emerging as a key paradigm to enhance it. However, RLVR training often suffers from policy…

Machine Learning · Computer Science 2026-04-20 Xiaoyun Zhang , Xiaojian Yuan , Di Huang , Wang You , Chen Hu , Jingqing Ruan , Ai Jian , Kejiang Chen , Xing Hu

Reinforcement Learning (RL) with PPO-like clip objectives has become the standard choice for reward-based fine-tuning of large language models (LLMs). Although recent work has explored improved estimators of advantages and normalization,…

Machine Learning · Computer Science 2026-02-24 Philipp Becker , Niklas Freymuth , Serge Thilges , Fabian Otto , Gerhard Neumann

Supervised Fine-Tuning (SFT) is the standard paradigm for domain adaptation, yet it frequently incurs the cost of catastrophic forgetting. In sharp contrast, on-policy Reinforcement Learning (RL) effectively preserves general capabilities.…

Machine Learning · Computer Science 2026-01-06 Muxi Diao , Lele Yang , Wuxuan Gong , Yutong Zhang , Zhonghao Yan , Yufei Han , Kongming Liang , Weiran Xu , Zhanyu Ma

Reinforcement Learning with Verifiable Rewards (RLVR) has catalyzed a leap in Large Language Model (LLM) reasoning, yet its optimization dynamics remain fragile. Standard algorithms like GRPO enforce stability via "hard clipping", which…

Machine Learning · Computer Science 2026-04-21 Xiaoliang Fu , Jiaye Lin , Yangyi Fang , Chaowen Hu , Cong Qin , Zekai Shao , Binbin Zheng , Lu Pan , Ke Zeng

Recent advances in rule-based reinforcement learning (RL) have significantly improved the reasoning capability of language models (LMs) with rule-based rewards. However, existing RL methods -- such as GRPO, REINFORCE++, and RLOO -- often…

Machine Learning · Computer Science 2025-05-20 Zongkai Liu , Fanqing Meng , Lingxiao Du , Zhixiang Zhou , Chao Yu , Wenqi Shao , Qiaosheng Zhang

Off-Policy reinforcement learning (RL) is an important class of methods for many problem domains, such as robotics, where the cost of collecting data is high and on-policy methods are consequently intractable. Standard methods for applying…

Artificial Intelligence · Computer Science 2019-07-03 Riley Simmons-Edler , Ben Eisner , Eric Mitchell , Sebastian Seung , Daniel Lee

Reinforcement Learning with Verified Reward (RLVR) has emerged as a critical paradigm for advancing the reasoning capabilities of Large Language Models (LLMs). Most existing RLVR methods, such as GRPO and its variants, ensure stable updates…

Machine Learning · Computer Science 2026-02-10 Qingyuan Wu , Yuhui Wang , Simon Sinong Zhan , Yanning Dai , Shilong Deng , Sarra Habchi , Qi Zhu , Matthias Gallé , Chao Huang

Existing training criteria in automatic speech recognition(ASR) permit the model to freely explore more than one time alignments between the feature and label sequences. In this paper, we use entropy to measure a model's uncertainty, i.e.…

Computation and Language · Computer Science 2022-12-26 Ehsan Variani , Ke Wu , David Rybach , Cyril Allauzen , Michael Riley

This dissertation investigates how reinforcement learning (RL) methods can be designed to be safe, sample-efficient, and robust. Framed through the unifying perspective of contextual-bandit RL, the work addresses two major application…

Machine Learning · Computer Science 2025-10-20 Shashank Gupta

To encourage diverse exploration in reinforcement learning (RL) for large language models (LLMs) without compromising accuracy, we propose Policy Split, a novel paradigm that bifurcates the policy into normal and high-entropy modes with a…

Computation and Language · Computer Science 2026-04-14 Jiashu Yao , Heyan Huang , Chuwei Luo , Daiqing Wu , Zeming Liu , Yuhang Guo , Yangyang Kang

Many of the challenges facing today's reinforcement learning (RL) algorithms, such as robustness, generalization, transfer, and computational efficiency are closely related to compression. Prior work has convincingly argued why minimizing…

Machine Learning · Computer Science 2021-09-08 Benjamin Eysenbach , Ruslan Salakhutdinov , Sergey Levine

We study risk-sensitive reinforcement learning in finite discounted MDPs with recursive entropic risk measures (ERM), where the risk parameter $\beta \neq 0$ controls the agent's risk attitude: $\beta>0$ for risk-averse and $\beta<0$ for…

Machine Learning · Computer Science 2026-05-20 Oliver Mortensen , Mohammad Sadegh Talebi

The safe application of reinforcement learning (RL) requires generalization from limited training data to unseen scenarios. Yet, fulfilling tasks under changing circumstances is a key challenge in RL. Current state-of-the-art approaches for…

Machine Learning · Computer Science 2023-12-06 Philipp Altmann , Fabian Ritz , Leonard Feuchtinger , Jonas Nüßlein , Claudia Linnhoff-Popien , Thomy Phan
‹ Prev 1 3 4 5 6 7 10 Next ›