AtomMem : Learnable Dynamic Agentic Memory with Atomic Memory Operation

TL;DR

AtomMem introduces a learnable, dynamic memory management framework for agents, transforming memory operations into a decision process optimized via supervised and reinforcement learning, leading to improved performance on long-context tasks.

Contribution

The paper presents AtomMem, a novel framework that learns to manage memory dynamically through atomic CRUD operations, enhancing flexibility and task alignment over static methods.

Findings

AtomMem-8B outperforms prior static memory methods on 3 benchmarks.

Learning-based memory management discovers structured, task-specific strategies.

Combining supervised and reinforcement learning improves memory orchestration.

Abstract

Equipping agents with memory is essential for solving real-world long-horizon problems. However, most existing agent memory mechanisms rely on static and hand-crafted workflows. This limits the performance and generalization ability of these memory designs, which highlights the need for a more flexible, learning-based memory framework. In this paper, we propose AtomMem, which reframes memory management as a dynamic decision-making problem. We deconstruct high-level memory processes into fundamental atomic CRUD (Create, Read, Update, Delete) operations, transforming the memory workflow into a learnable decision process. By combining supervised fine-tuning with reinforcement learning, AtomMem learns an autonomous, task-aligned policy to orchestrate memory behaviors tailored to specific task demands. Experimental results across 3 long-context benchmarks demonstrate that the trained AtomMem-8B consistently outperforms prior static-workflow memory methods. Further analysis of training dynamics shows that our learning-based formulation enables the agent to discover structured, task-aligned memory management strategies, highlighting a key advantage over predefined routines.