Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | https://arxiv.org/abs/2604.00131 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910141615439872 |
|---|---|
| author | Rana, Ashish Hung, Chia-Chien Sun, Qumeng Kunkel, Julian Martin Lawrence, Carolin |
| author_facet | Rana, Ashish Hung, Chia-Chien Sun, Qumeng Kunkel, Julian Martin Lawrence, Carolin |
| contents | Human memory adapts through selective forgetting: experiences become less accessible over time but can be reactivated by reinforcement or contextual cues. In contrast, memory-augmented LLM agents rely on "always-on" retrieval and "flat" memory storage, causing high interference and latency as histories grow. We introduce Oblivion, a memory control framework that casts forgetting as decay-driven reductions in accessibility, not explicit deletion. Oblivion decouples memory control into read and write paths. The read path decides when to consult memory, based on agent uncertainty and memory buffer sufficiency, avoiding redundant always-on access. The write path decides what to strengthen, by reinforcing memories contributing to forming the response. Together, this enables hierarchical memory organization that maintains persistent high-level strategies while dynamically loading details as needed. We evaluate on both static and dynamic long-horizon interaction benchmarks. Results show that Oblivion dynamically adapts memory access and reinforcement, balancing learning and forgetting under shifting contexts, highlighting that memory control is essential for effective LLM-agentic reasoning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_00131 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Oblivion: Self-Adaptive Agentic Memory Control through Decay-Driven Activation Rana, Ashish Hung, Chia-Chien Sun, Qumeng Kunkel, Julian Martin Lawrence, Carolin Computation and Language Artificial Intelligence Human memory adapts through selective forgetting: experiences become less accessible over time but can be reactivated by reinforcement or contextual cues. In contrast, memory-augmented LLM agents rely on "always-on" retrieval and "flat" memory storage, causing high interference and latency as histories grow. We introduce Oblivion, a memory control framework that casts forgetting as decay-driven reductions in accessibility, not explicit deletion. Oblivion decouples memory control into read and write paths. The read path decides when to consult memory, based on agent uncertainty and memory buffer sufficiency, avoiding redundant always-on access. The write path decides what to strengthen, by reinforcing memories contributing to forming the response. Together, this enables hierarchical memory organization that maintains persistent high-level strategies while dynamically loading details as needed. We evaluate on both static and dynamic long-horizon interaction benchmarks. Results show that Oblivion dynamically adapts memory access and reinforcement, balancing learning and forgetting under shifting contexts, highlighting that memory control is essential for effective LLM-agentic reasoning. |
| title | Oblivion: Self-Adaptive Agentic Memory Control through Decay-Driven Activation |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2604.00131 |