作者位次 #3/11ACL 2026Poster· Natural Language Processing
Learning How and What to Memorize: Cognition-Inspired Two-Stage Optimization for Evolving Memory
Derong Xu · Shuochen Liu · Pengfei Luo · an Jiaotong University) · Yingyi Zhang · Yi Wen · Yimin Deng · Wenlin Zhang · Enhong Chen (University of Science and Technology of China & · an Jiaotong University) · an Jiaotong University)
摘要
Large language model (LLM) agents require long-term user memory for consistent personalization, but limited context windows hinder tracking evolving preferences over long interactions. Existing memory systems mainly rely on static, hand-crafted update rules; although reinforcement learning (RL)-b...