HAGE: Harnessing Agentic Memory via RL-Driven Weighted Graph Evolution

2026-05-10

AI Summary

This work addresses the limitations of existing agent memory systems, which often rely on static vector retrieval or fixed binary relation graphs and thus struggle to capture the strength, confidence, and query relevance of event relationships. To overcome these challenges, the authors propose a learnable weighted multi-relational memory graph that organizes memories as a multi-view graph over shared nodes. The framework incorporates query-aware edge embedding modulation and a dynamic routing mechanism, enabling sequential and adaptive memory retrieval. Edge representations and traversal policies are jointly optimized via reinforcement learning. Evaluated on long-horizon reasoning tasks, the proposed method achieves significantly higher accuracy and demonstrates superior trade-offs between accuracy and efficiency compared to current agent memory architectures.

Abstract

Memory retrieval in agentic large language model (LLM) systems is often treated as a static lookup problem, relying on flat vector search or fixed binary relational graphs. However, fixed graph structures cannot capture the varying strength, confidence, and query-dependent relevance of relationships between events. In this paper, we propose HAGE, a weighted multi-relational memory framework that reconceptualizes retrieval as sequential, query-conditioned traversal over a unified relational memory graph. Memory is organized as relation-specific graph views over shared memory nodes, where each edge is associated with a trainable relation feature vector encoding multiple relational signals. Given a query, an LLM-based classifier identifies the relational intent, and a routing network dynamically modulates the corresponding dimensions of the edge embedding. Traversal scores are computed via a learned combination of semantic similarity and these query-conditioned edge representations. This allows memory traversal to prioritize high-utility relational paths while softly suppressing noisy or weakly relevant connections. Beyond adaptive traversal, HAGE further introduces a reinforcement learning-based training framework that jointly optimizes routing behavior and edge representations using downstream tasks. Finally, empirical results demonstrate improved long-horizon reasoning accuracy and a favorable accuracy-efficiency trade-off compared to state-of-the-art agentic memory systems. Our code is available at https://github.com/FredJiang0324/HAGE_MVPReview.

Problem

• memory retrieval
• agentic LLM
• relational graph
• query-dependent relevance
• static lookup

Innovation

• weighted relational graph
• query-conditioned traversal
• reinforcement learning
• agentic memory
• dynamic routing