long-term-memory · activeRemembering across sessions
Assistants fail to recall and apply facts, preferences, and decisions from earlier sessions.
Also called: persistent memory, cross-session memory, personalization
Tags: autonomous-agent, chat-assistant, context, customer-support
A capable assistant carries forward what a user told it and what was decided, notices when a stored fact has changed, and reasons over information spread across many past interactions.
Claims
- mechanismsingle paperCommercial assistants show a large accuracy drop when the relevant information sits in a long interaction history, especially for updates and multi-session reasoning.
- mechanismsingle paperA memory stream with retrieval and reflection lets agents behave consistently over simulated days.
- mechanismsingle paperFor persistent-memory agents, any memory-write rule that decides using only recency and provenance is stuck on a single tradeoff — accepting more genuine preference updates means admitting more poisoned ones — because an adversary who can launder a claim through the user's own channel matches the statistics a genuine revision produces; conditioning the write decision on the inferred authenticity of the claim given the interaction history moves both axes at once.
- observationsingle paperWhen the preference needed to answer a query lies in a different concept space from the cues in a user's dialogue history (e.g. inferring an emotion-regulation strategy from personality cues), retrieval-based and profile-summarization personalization methods barely beat a no-history baseline, and dumping the full history into the prompt does worse than no history at all.
- mechanismmechanism reasoningA running summary buys continuity across a long conversation at the cost of specifics, because it is lossy compression by construction — so it helps with "what were we doing" and not with "what exactly did I say in turn 40." Keeping the raw turns retrievable alongside the summary covers the second case.
- mechanismsingle paperRecursively summarizing the dialogue into a running memory improves consistency in long conversations.
- observationsingle paperWhen a structured long-term memory substrate (indexes, retrieval, control logic) is held fixed and only the backbone LLM is swapped, accuracy on multi-session memory benchmarks like LongMemEval and LoCoMo moves by a few points while per-query cost varies by roughly an order of magnitude, so recall quality in these settings is set mostly by the memory and retrieval design rather than by model choice.
- mechanismsingle paperA tiered memory system managed by the model itself sustains recall over conversations far longer than the context window.
Techniques
- Checkpoint task state to a fileprocessWrite task id, status, completed steps and artifact paths to a small file after every meaningful step, and read it at session start.
- Maintain a running summarypromptingPeriodically fold the conversation into a summary that replaces older turns.
- Tiered memory managed by the modelarchitectureKeep a small working context and let the model page facts in and out of external storage.
Capabilities are a way of carving up the subject, and carvings are arguable. Say so if this one is wrong — especially a proposed one, which a pipeline added because several papers used the same framing, not because anyone decided it was right.