The Landscape of Agentic Reinforcement Learning for LLMs: A Survey
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866915942591627264 |
|---|---|
| author | Zhang, Guibin Geng, Hejia Yu, Xiaohang Yin, Zhenfei Zhang, Zaibin Tan, Zelin Zhou, Heng Li, Zhongzhi Xue, Xiangyuan Li, Yijiang Zhou, Yifan Chen, Yang Zhang, Chen Fan, Yutao Wang, Zihu Huang, Songtao Piedrahita-Velez, Francisco Liao, Yue Wang, Hongru Yang, Mengyue Ji, Heng Wang, Jun Yan, Shuicheng Torr, Philip Bai, Lei |
| author_facet | Zhang, Guibin Geng, Hejia Yu, Xiaohang Yin, Zhenfei Zhang, Zaibin Tan, Zelin Zhou, Heng Li, Zhongzhi Xue, Xiangyuan Li, Yijiang Zhou, Yifan Chen, Yang Zhang, Chen Fan, Yutao Wang, Zihu Huang, Songtao Piedrahita-Velez, Francisco Liao, Yue Wang, Hongru Yang, Mengyue Ji, Heng Wang, Jun Yan, Shuicheng Torr, Philip Bai, Lei |
| contents | The emergence of agentic reinforcement learning (Agentic RL) marks a paradigm shift from conventional reinforcement learning applied to large language models (LLM RL), reframing LLMs from passive sequence generators into autonomous, decision-making agents embedded in complex, dynamic worlds. This survey formalizes this conceptual shift by contrasting the degenerate single-step Markov Decision Processes (MDPs) of LLM-RL with the temporally extended, partially observable Markov decision processes (POMDPs) that define Agentic RL. Building on this foundation, we propose a comprehensive twofold taxonomy: one organized around core agentic capabilities, including planning, tool use, memory, reasoning, self-improvement, and perception, and the other around their applications across diverse task domains. Central to our thesis is that reinforcement learning serves as the critical mechanism for transforming these capabilities from static, heuristic modules into adaptive, robust agentic behavior. To support and accelerate future research, we consolidate the landscape of open-source environments, benchmarks, and frameworks into a practical compendium. By synthesizing over five hundred recent works, this survey charts the contours of this rapidly evolving field and highlights the opportunities and challenges that will shape the development of scalable, general-purpose AI agents. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_02547 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | The Landscape of Agentic Reinforcement Learning for LLMs: A Survey Zhang, Guibin Geng, Hejia Yu, Xiaohang Yin, Zhenfei Zhang, Zaibin Tan, Zelin Zhou, Heng Li, Zhongzhi Xue, Xiangyuan Li, Yijiang Zhou, Yifan Chen, Yang Zhang, Chen Fan, Yutao Wang, Zihu Huang, Songtao Piedrahita-Velez, Francisco Liao, Yue Wang, Hongru Yang, Mengyue Ji, Heng Wang, Jun Yan, Shuicheng Torr, Philip Bai, Lei Artificial Intelligence Computation and Language The emergence of agentic reinforcement learning (Agentic RL) marks a paradigm shift from conventional reinforcement learning applied to large language models (LLM RL), reframing LLMs from passive sequence generators into autonomous, decision-making agents embedded in complex, dynamic worlds. This survey formalizes this conceptual shift by contrasting the degenerate single-step Markov Decision Processes (MDPs) of LLM-RL with the temporally extended, partially observable Markov decision processes (POMDPs) that define Agentic RL. Building on this foundation, we propose a comprehensive twofold taxonomy: one organized around core agentic capabilities, including planning, tool use, memory, reasoning, self-improvement, and perception, and the other around their applications across diverse task domains. Central to our thesis is that reinforcement learning serves as the critical mechanism for transforming these capabilities from static, heuristic modules into adaptive, robust agentic behavior. To support and accelerate future research, we consolidate the landscape of open-source environments, benchmarks, and frameworks into a practical compendium. By synthesizing over five hundred recent works, this survey charts the contours of this rapidly evolving field and highlights the opportunities and challenges that will shape the development of scalable, general-purpose AI agents. |
| title | The Landscape of Agentic Reinforcement Learning for LLMs: A Survey |
| topic | Artificial Intelligence Computation and Language |
| url | https://arxiv.org/abs/2509.02547 |