All-Mem: Agentic Lifelong Memory via Dynamic Topology Evolution
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866908902233210880 |
|---|---|
| author | Lv, Can Chang, Heng Guo, Yuchen Tao, Shengyu Zhou, Shiji |
| author_facet | Lv, Can Chang, Heng Guo, Yuchen Tao, Shengyu Zhou, Shiji |
| contents | Lifelong interactive agents are expected to assist users over months or years, which requires continually writing long term memories while retrieving the right evidence for each new query under fixed context and latency budgets. Existing memory systems often degrade as histories grow, yielding redundant, outdated, or noisy retrieved contexts. We present All-Mem, an online/offline lifelong memory framework that maintains a topology structured memory bank via explicit, non destructive consolidation, avoiding the irreversible information loss typical of summarization based compression. In online operation, it anchors retrieval on a bounded visible surface to keep coarse search cost bounded. Periodically offline, an LLM diagnoser proposes confidence scored topology edits executed with gating using three operators: SPLIT, MERGE, and UPDATE, while preserving immutable evidence for traceability. At query time, typed links enable hop bounded, budgeted expansion from active anchors to archived evidence when needed. Experiments on LOCOMO and LONGMEMEVAL show improved retrieval and QA over representative baselines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_19595 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | All-Mem: Agentic Lifelong Memory via Dynamic Topology Evolution Lv, Can Chang, Heng Guo, Yuchen Tao, Shengyu Zhou, Shiji Information Retrieval Computation and Language Lifelong interactive agents are expected to assist users over months or years, which requires continually writing long term memories while retrieving the right evidence for each new query under fixed context and latency budgets. Existing memory systems often degrade as histories grow, yielding redundant, outdated, or noisy retrieved contexts. We present All-Mem, an online/offline lifelong memory framework that maintains a topology structured memory bank via explicit, non destructive consolidation, avoiding the irreversible information loss typical of summarization based compression. In online operation, it anchors retrieval on a bounded visible surface to keep coarse search cost bounded. Periodically offline, an LLM diagnoser proposes confidence scored topology edits executed with gating using three operators: SPLIT, MERGE, and UPDATE, while preserving immutable evidence for traceability. At query time, typed links enable hop bounded, budgeted expansion from active anchors to archived evidence when needed. Experiments on LOCOMO and LONGMEMEVAL show improved retrieval and QA over representative baselines. |
| title | All-Mem: Agentic Lifelong Memory via Dynamic Topology Evolution |
| topic | Information Retrieval Computation and Language |
| url | https://arxiv.org/abs/2603.19595 |