analiza aiqbench
This research addresses a critical bottleneck in training LLM agents by demonstrating that dense, per-turn reward structures significantly outperform sparse trajectory-level rewards for fine-grained credit assignment. The study provides tailored implementations of GRPO and PPO algorithms that achieve faster convergence and higher accuracy in complex multi-turn search and game tasks. For practitioners building autonomous agents, this work offers a concrete methodological shift toward leveraging intermediate feedback to stabilize long-horizon reasoning.
Reinforcing Multi-Turn Reasoning in LLM Agents via Fine-Grained Reward Structure and Credit Assignment
arXiv:2505.11821v3 Announce Type: replace Abstract: Reinforcement Learning (RL) approaches have been wildly used to enhance the reasoning capabilities of Large Language Model (LLM) agents in long-horizon, multi-turn scenarios. Such interactions can be formalized as turn-level Markov decision processes (MDPs), where intermediate rewards are often available. However, most prior work relies on sparse trajectory-level rewards, resulting in poor credit assignment, while dense turn-level rewards remain underexplored. In this paper, we investigate how to effectively leverage dense turn-level reward structures in RL algorithms, specifically Group Relative Policy Optimization (GRPO) and Proximal Policy Optimization (PPO), to enable fine-grained credit assignment. We categorize reward structures into three types based on their granularity: (1) terminal reward; (2) delayed reward; (3) per-turn reward, each corresponding to a distinct turn-level MDP formulation, and derive GRPO and PPO algorithms tailored to each case, respectively. Experiments on multi-turn search and game agent tasks show that, for both GRPO and PPO, dense per-turn reward structures consistently outperform sparse terminal and delayed reward structures in terms of training dynamics and numerical results. Furthermore, on search tasks, PPO with dense per-turn rewards achieves greater training stability and faster convergence, and delivers the highest answer correctness among all state-of-the-art methods across diverse question-answering datasets.
Abstract:Reinforcement Learning (RL) approaches have been wildly used to enhance the reasoning capabilities of Large Language Model (LLM) agents in long-horizon, multi-turn scenarios. Such interactions can be formalized as turn-level Markov decision processes (MDPs), where intermediate rewards are often available. However, most prior work relies on sparse trajectory-level rewards, resulting in poor credit assignment, while dense turn-level rewards remain underexplored. In this paper, we investigate how to effectively leverage dense turn-level reward structures in RL algorithms, specifically Group Relative Policy Optimization (GRPO) and Proximal Policy Optimization (PPO), to enable fine-grained credit assignment. We categorize reward structures into three types based on their granularity: (1) terminal reward; (2) delayed reward; (3) per-turn reward, each corresponding to a distinct turn-level MDP formulation, and derive GRPO and PPO algorithms tailored to each case, respectively. Experiments on multi-turn search and game agent tasks show that, for both GRPO and PPO, dense per-turn reward structures consistently outperform sparse terminal and delayed reward structures in terms of training dynamics and numerical results. Furthermore, on search tasks, PPO with dense per-turn rewards achieves greater training stability and faster convergence, and delivers the highest answer correctness among all state-of-the-art methods across diverse question-answering datasets.
Machine Learning (cs.LG)
arXiv:2505.11821 [cs.LG]
(or arXiv:2505.11821v3 [cs.LG] for this version)
https://doi.org/10.48550/arXiv.2505.11821 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 288 cuvinte
Subiecte conexe
ARISE-RL: Agentic Rubric-Grounded Iterative Self-Evolution with Reinforcement Learning
arXiv · 2 sept. 2026
Introducing North Mini Code: Cohere’s First Model For Developers
Hugging Face · 9 iun. 2026
What's Continual Learning, and Why Might We Expect To See It In Advanced LLM Agents?
LessWrong · 12 iun. 2026
PBSD: Privileged Bayesian Self-Distillation for Long-Horizon Credit Assignment
arXiv · 9 iun. 2026