Skip to content

Source paper

Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolution

Zouying Cao, Jiaji Deng, Li Yu and 4 more

Zouying Cao, Jiaji Deng, Li Yu, Weikang Zhou, Zhaoyang Liu, Bolin Ding, Hai Zhao

arXiv 2512.10696 · 2025

Research ideas from this paper (4)

Generated by AI. The paper’s authors did not write these.

Dynamic Multi-Step Retrieval for Long-Horizon Tasks

Empower long-horizon agents to query memory autonomously during task execution by training them to emit an active retrieval token when confidence drops.

Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolution
🟠Medium💡Moderate InnovationLarge clusterRelated work found#Retrieval & memory#Agents

Simulation-Based Ground Truth Validation

Validate extracted agent experiences objectively by running temporary sandbox simulations to verify whether candidate memories actually improve task success.

Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolution
🔴Hard🔬Significant ResearchMultiple GPUsNo close match found#Data & evaluation#Agents

Distilled 'Memory Cortex' for Efficient Self-Evolution

Distill high-performance summarization capabilities into a lightweight memory model by training a small architecture on outputs from a larger teacher model.

Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolution
🟡Fairly easy💡Moderate InnovationLarge clusterRelated work found#Retrieval & memory#LLMs

Semantic Density Pruning for Memory Optimization

Optimize agent memory pools by clustering semantically similar experiences and using an LLM to merge them into unified canonical representations.

Remember Me, Refine Me: A Dynamic Procedural Memory Framework for Experience-Driven Agent Evolution
🟢Easy🔧Minor TweakMultiple GPUsRelated work found#Retrieval & memory#Efficiency

AI summary

This paper introduces ReMe (Remember Me, Refine Me), a framework for autonomous agent evolution that replaces static memory accumulation with a dynamic lifecycle of experience extraction, reuse, and refinement. By distilling keypoint-level insights from trajectories and employing utility-based pruning, ReMe enables smaller models to outperform larger baselines on complex reasoning tasks without parameter updates.

How it works (AI summary)

ReMe operates in three phases: 1) Experience Acquisition, where an LLM summarizer distills success patterns and failure analysis from trajectories into structured keypoint experiences; 2) Experience Reuse, which retrieves, reranks, and rewrites experiences based on usage scenarios to guide the agent; and 3) Experience Refinement, which manages the memory pool by selectively adding valid insights and deleting low-utility entries.