AtomMem : Learnable Dynamic Agentic Memory with Atomic Memory Operation
About
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.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Open-domain Question Answering | LoCoMo Open-Domain | F1 Score6.37 | 24 | |
| Single-hop Question Answering | LoCoMo Single-Hop | F1 Score9.72 | 24 | |
| Question Answering | LoCoMo Multi-Hop | F1 Score8.87 | 20 | |
| Memory Extraction | HaluMem | Memory Accuracy87.5 | 19 | |
| Memory Updating | HaluMem | C Score2.14 | 17 | |
| Question Answering | HaluMem | Metric C31.27 | 17 | |
| Long-context memory management | MemoryAgentBench | Single-Doc Recall37 | 16 | |
| Long-term Temporal Profile evaluation | LTP (Long-term Temporal Profile) 10 profiles (800 judged probes) (Full) | Bank Coexist100 | 12 | |
| Question Answering | LoCoMo Adversarial | F1 Score4.48 | 12 | |
| Question Answering | LoCoMo Average scores | F1 Score6.72 | 12 |