MemComposer: Learning to Assemble Memories for LLM Agents via Reinforcement Learning

Key points

• However, semantic similarity alone does not guarantee memory usefulness: downstream performance depends on how memories are selected and presented to the agent.1
• We propose MemComposer, a memory composer that jointly examines the current query and retrieved candidates to decide which experiences to include, how many to select, and how to order them.1
• We train the composer using Experience Policy Optimization (ExPO), a GRPO-based framework that evaluates assembled memory contexts through complete agent interactions and uses task-success feedback to optimize the assembly policy.1
• These results demonstrate the value of learning query-specific memory assembly using downstream task feedback.1

Sources (1)

• [1]MemComposer: Learning to Assemble Memories for LLM Agents via Reinforcement Learning
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 15, 01:47 AM

  Agent memory enables large language model (LLM) agents to reuse past experiences without updating their parameters.

  However, semantic similarity alone does not guarantee memory usefulness: downstream performance depends on how memories are selected and presented to the agent.

Extractive summary: sentences quoted from the sources.