A LoRA memory system lets stale user facts fade inside model weights
A proposed memory system lets an LLM retain stable user facts, update changed beliefs, and let temporary information expire without relying only on retrieval. It stores facts in a fixed-capacity LoRA adapter, then uses a lightweight controller to decide what to remember, how deeply, and for how long. On long-horizon user trajectories, the policy reached 86% of oracle performance while testing both recall during validity and abstention after expiry.