Contextual Experience Replay for Self-Improvement of Language Agents

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Liu, Yitao, Si, Chenglei, Narasimhan, Karthik, Yao, Shunyu
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866909642360094720
author Liu, Yitao
Si, Chenglei
Narasimhan, Karthik
Yao, Shunyu
author_facet Liu, Yitao
Si, Chenglei
Narasimhan, Karthik
Yao, Shunyu
contents Large language model (LLM) agents have been applied to sequential decision-making tasks such as web navigation, but without any environment-specific experiences, they often fail in these complex tasks. Moreover, current LLM agents are not designed to continually learn from past experiences during inference time, which could be crucial for them to gain these environment-specific experiences. To address this, we propose Contextual Experience Replay (CER), a training-free framework to enable efficient self-improvement for language agents in their context window. Specifically, CER accumulates and synthesizes past experiences into a dynamic memory buffer. These experiences encompass environment dynamics and common decision-making patterns, allowing the agents to retrieve and augment themselves with relevant knowledge in new tasks, enhancing their adaptability in complex environments. We evaluate CER on the challenging WebArena and VisualWebArena benchmarks. On VisualWebArena, CER achieves a competitive performance of 31.9%. On WebArena, CER also gets a competitive average success rate of 36.7%, relatively improving the success rate of the GPT-4o agent baseline by 51.0%. We also conduct a comprehensive analysis on it to prove its efficiency, validity and understand it better.
format Preprint
id arxiv_https___arxiv_org_abs_2506_06698
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Contextual Experience Replay for Self-Improvement of Language Agents
Liu, Yitao
Si, Chenglei
Narasimhan, Karthik
Yao, Shunyu
Artificial Intelligence
Computation and Language
Computer Vision and Pattern Recognition
Machine Learning
Large language model (LLM) agents have been applied to sequential decision-making tasks such as web navigation, but without any environment-specific experiences, they often fail in these complex tasks. Moreover, current LLM agents are not designed to continually learn from past experiences during inference time, which could be crucial for them to gain these environment-specific experiences. To address this, we propose Contextual Experience Replay (CER), a training-free framework to enable efficient self-improvement for language agents in their context window. Specifically, CER accumulates and synthesizes past experiences into a dynamic memory buffer. These experiences encompass environment dynamics and common decision-making patterns, allowing the agents to retrieve and augment themselves with relevant knowledge in new tasks, enhancing their adaptability in complex environments. We evaluate CER on the challenging WebArena and VisualWebArena benchmarks. On VisualWebArena, CER achieves a competitive performance of 31.9%. On WebArena, CER also gets a competitive average success rate of 36.7%, relatively improving the success rate of the GPT-4o agent baseline by 51.0%. We also conduct a comprehensive analysis on it to prove its efficiency, validity and understand it better.
title Contextual Experience Replay for Self-Improvement of Language Agents
topic Artificial Intelligence
Computation and Language
Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2506.06698