作者位次 #2/5ACL 2026Findings· Findings of ACL
From Recall to Forgetting: Benchmarking Long-Term Memory for Personalized Agents
Md Nayem Uddin · Kumar Shubham · Eduardo Blanco · Chitta Baral · Gengyu Wang
摘要
Personalized agents that interact with users over long periods must maintain persistent memory across sessions and update it as circumstances change. However, existing benchmarks predominantly frame long-term memory evaluation as fact retrieval from past conversations, providing limited insight i...