Memp: Exploring Agent Procedural Memory

ACL ARR 2026 January Submission7841 Authors

Keywords: LLM/AI agents, procedural memory, life-long learning

Abstract: Large Language Models (LLMs) based agents excel at diverse tasks, yet they suffer from brittle procedural memory that is manually engineered or entangled in static parameters. In this work, we investigate strategies to endow agents with a learnable, updatable, and lifelong procedural memory. We propose a procedural-memory repository that distills past agent trajectories into both fine-grained, step-by-step instructions and higher-level, script-like abstractions. Coupled with a dynamic regimen that continuously updates, corrects, and deprecates its contents, this repository evolves in lockstep with new experience. Empirical evaluation on TravelPlanner and Alfworld shows that as the memory repository is refined, agents achieve steadily higher success rates and greater efficiency on analogous tasks. Moreover, procedural memory built from a stronger model retains its value: migrating the procedural memory to a weaker model yields substantial performance gains.

Paper Type: Long

Research Area: AI/LLM Agents

Research Area Keywords: LLM/AI agents, prompting

Contribution Types: NLP engineering experiment

Languages Studied: English

Submission Number: 7841