BMAM: Brain-inspired Multi-Agent Memory Framework
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866917227344691200 |
|---|---|
| author | Li, Yang Liu, Jiaxiang Wang, Yusong Wu, Yujie Xu, Mingkun |
| author_facet | Li, Yang Liu, Jiaxiang Wang, Yusong Wu, Yujie Xu, Mingkun |
| contents | Language-model-based agents operating over extended interaction horizons face persistent challenges in preserving temporally grounded information and maintaining behavioral consistency across sessions, a failure mode we term soul erosion. We present BMAM (Brain-inspired Multi-Agent Memory), a general-purpose memory architecture that models agent memory as a set of functionally specialized subsystems rather than a single unstructured store. Inspired by cognitive memory systems, BMAM decomposes memory into episodic, semantic, salience-aware, and control-oriented components that operate at complementary time scales. To support long-horizon reasoning, BMAM organizes episodic memories along explicit timelines and retrieves evidence by fusing multiple complementary signals. Experiments on the LoCoMo benchmark show that BMAM achieves 78.45 percent accuracy under the standard long-horizon evaluation setting, and ablation analyses confirm that the hippocampus-inspired episodic memory subsystem plays a critical role in temporal reasoning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_20465 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | BMAM: Brain-inspired Multi-Agent Memory Framework Li, Yang Liu, Jiaxiang Wang, Yusong Wu, Yujie Xu, Mingkun Computation and Language 68T50, 68T05 I.2.7; I.2.4; H.3.3 Language-model-based agents operating over extended interaction horizons face persistent challenges in preserving temporally grounded information and maintaining behavioral consistency across sessions, a failure mode we term soul erosion. We present BMAM (Brain-inspired Multi-Agent Memory), a general-purpose memory architecture that models agent memory as a set of functionally specialized subsystems rather than a single unstructured store. Inspired by cognitive memory systems, BMAM decomposes memory into episodic, semantic, salience-aware, and control-oriented components that operate at complementary time scales. To support long-horizon reasoning, BMAM organizes episodic memories along explicit timelines and retrieves evidence by fusing multiple complementary signals. Experiments on the LoCoMo benchmark show that BMAM achieves 78.45 percent accuracy under the standard long-horizon evaluation setting, and ablation analyses confirm that the hippocampus-inspired episodic memory subsystem plays a critical role in temporal reasoning. |
| title | BMAM: Brain-inspired Multi-Agent Memory Framework |
| topic | Computation and Language 68T50, 68T05 I.2.7; I.2.4; H.3.3 |
| url | https://arxiv.org/abs/2601.20465 |