Experience Memory Graph: One-Shot Error Correction for Agents

2026-07-15

AI Summary

This work addresses the challenge of error accumulation in large language model agents during complex, long-horizon tasks, where existing self-reflection methods suffer from high computational costs due to repeated trial-and-error and limited generalization. The paper introduces a novel approach that formulates failure recovery as a graph matching problem, proposing a directed action decision graph-based memory structure to store past experiences. By leveraging subgraph matching and graph edit path analysis, the method extracts transferable corrective knowledge across tasks, enabling one-step, non-iterative error correction without requiring test-time trials. Experimental results on ALFWorld and ScienceWorld demonstrate that this approach significantly outperforms current reflection mechanisms, achieving higher success rates and average rewards while eliminating the need for in-situ exploration during inference.

Abstract

Large Language Model (LLM) agents have shown remarkable capabilities in autonomous decision-making by generating sequential trajectories of states, actions, and observations. However, in complex, long-horizon tasks, these agents frequently suffer from compounding errors and struggle to recover from failures. Existing self-correction mechanisms rely on prompt-based reflection, which is inherently brittle, incurs heavy time and API costs due to iterative trial-and-error loops, and produces task-specific memory that may be hard to generalize to new scenarios. To address this, we propose Experience Memory Graph (EMG), a framework that reformulates agent failure recovery as a graph matching problem. At training time, we convert both failed exploration trajectories and successful expert trajectories into directed action decision graphs. By matching these graphs, we extract common subgraphs (successful workflows) and graph edit paths that explicitly indicate how to correct failures (e.g., which actions to add, delete, or relabel under a given observation), and store them in a memory graph with intra-task nodes and cross-task edges. At test time, EMG retrieves relevant insights and guides the agent in a single, loop-free execution. Experiments on ALFWorld and ScienceWorld show that EMG consistently outperforms state-of-the-art reflection baselines in success rate and average reward, while requiring no test-time trial-and-error.

Problem

• error correction
• LLM agents
• failure recovery
• compounding errors
• generalization

Innovation

• Experience Memory Graph
• graph matching
• one-shot error correction
• action decision graph
• LLM agents