EDGE: Experience-Distillation for Guided Exploration in Agentic Reinforcement Learning
EDGE:智能体强化学习中引导探索的经验蒸馏方法
机构 * School of Artificial Intelligence, University of Chinese Academy of Sciences(中国科学院大学人工智能学院) ; Institute of Automation, Chinese Academy of Sciences(中国科学院自动化研究所) ; School of Advanced Interdisciplinary Sciences, University of Chinese Academy of Sciences(中国科学院大学先进交叉科学学院) ; Wuhan AI Research(武汉人工智能研究院)
AI总结 本文提出EDGE框架,将检索的经验内化到策略中,在ALFWorld和WebShop数据集7B规模下较GRPO提升8.3、12.5个成功率百分点,移除外部经验后仍保留96.0%支架性能。
Comments Accepted to EMNLP 2026 (Main Conference)