UMEM: Unified Memory Extraction and Management Framework for Generalizable Memory
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911440018866176 |
|---|---|
| author | Ye, Yongshi Jiang, Hui Jiang, Feihu Lan, Tian Du, Yichao Fu, Biao Shi, Xiaodong Jia, Qianghuai Wang, Longyue Luo, Weihua |
| author_facet | Ye, Yongshi Jiang, Hui Jiang, Feihu Lan, Tian Du, Yichao Fu, Biao Shi, Xiaodong Jia, Qianghuai Wang, Longyue Luo, Weihua |
| contents | Self-evolving memory serves as the trainable parameters for Large Language Models (LLMs)-based agents, where extraction (distilling insights from experience) and management (updating the memory bank) must be tightly coordinated. Existing methods predominately optimize memory management while treating memory extraction as a static process, resulting in poor generalization, where agents accumulate instance-specific noise rather than robust memories. To address this, we propose Unified Memory Extraction and Management (UMEM), a self-evolving agent framework that jointly optimizes a Large Language Model to simultaneous extract and manage memories. To mitigate overfitting to specific instances, we introduce Semantic Neighborhood Modeling and optimize the model with a neighborhood-level marginal utility reward via GRPO. This approach ensures memory generalizability by evaluating memory utility across clusters of semantically related queries. Extensive experiments across five benchmarks demonstrate that UMEM significantly outperforms highly competitive baselines, achieving up to a 10.67% improvement in multi-turn interactive tasks. Futhermore, UMEM maintains a monotonic growth curve during continuous evolution. Codes and models will be publicly released. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_10652 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | UMEM: Unified Memory Extraction and Management Framework for Generalizable Memory Ye, Yongshi Jiang, Hui Jiang, Feihu Lan, Tian Du, Yichao Fu, Biao Shi, Xiaodong Jia, Qianghuai Wang, Longyue Luo, Weihua Computation and Language Self-evolving memory serves as the trainable parameters for Large Language Models (LLMs)-based agents, where extraction (distilling insights from experience) and management (updating the memory bank) must be tightly coordinated. Existing methods predominately optimize memory management while treating memory extraction as a static process, resulting in poor generalization, where agents accumulate instance-specific noise rather than robust memories. To address this, we propose Unified Memory Extraction and Management (UMEM), a self-evolving agent framework that jointly optimizes a Large Language Model to simultaneous extract and manage memories. To mitigate overfitting to specific instances, we introduce Semantic Neighborhood Modeling and optimize the model with a neighborhood-level marginal utility reward via GRPO. This approach ensures memory generalizability by evaluating memory utility across clusters of semantically related queries. Extensive experiments across five benchmarks demonstrate that UMEM significantly outperforms highly competitive baselines, achieving up to a 10.67% improvement in multi-turn interactive tasks. Futhermore, UMEM maintains a monotonic growth curve during continuous evolution. Codes and models will be publicly released. |
| title | UMEM: Unified Memory Extraction and Management Framework for Generalizable Memory |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2602.10652 |