Beyond Static Summarization: Proactive Memory Extraction for LLM Agents

2026-01-08

arXiv.org

AI Summary

Current memory retrieval approaches for large language model agents predominantly rely on static, single-pass summarization, which lacks awareness of task-specific requirements and factual verification, often leading to critical information loss and error propagation. This work proposes ProMem, an active memory retrieval framework that transcends the conventional “retrieve-once-and-early” paradigm by modeling memory access as a task-driven, iterative cognitive process. Through a self-querying mechanism, ProMem establishes a recursive feedback loop to dynamically inspect dialogue history, enabling recovery of missing information and correction of inaccuracies. Experimental results demonstrate that ProMem substantially enhances memory completeness and question-answering accuracy while achieving a more favorable trade-off between retrieval quality and token consumption.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Machine Learning: Active Learning
• Natural Language Processing: Summarization

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Memory management is vital for LLM agents to handle long-term interaction and personalization. Most research focuses on how to organize and use memory summary, but often overlooks the initial memory extraction stage. In this paper, we argue that existing summary-based methods have two major limitations based on the recurrent processing theory. First, summarization is"ahead-of-time", acting as a blind"feed-forward"process that misses important details because it doesn't know future tasks. Second, extraction is usually"one-off", lacking a feedback loop to verify facts, which leads to the accumulation of information loss. To address these issues, we propose proactive memory extraction (namely ProMem). Unlike static summarization, ProMem treats extraction as an iterative cognitive process. We introduce a recurrent feedback loop where the agent uses self-questioning to actively probe the dialogue history. This mechanism allows the agent to recover missing information and correct errors. Our ProMem significantly improves the completeness of the extracted memory and QA accuracy. It also achieves a superior trade-off between extraction quality and token cost.

Problem

• memory extraction
• LLM agents
• static summarization
• information loss
• feedback loop

Innovation

• proactive memory extraction
• recurrent feedback loop
• self-questioning
• LLM agents
• memory summarization