{"slug":"memory-long-term","label":"Memory Long-Term","item_count":3,"day_count":3,"source_count":2,"first_seen":"2026-08-17T09:14:31+00:00","last_updated":"2026-08-31T09:41:26+00:00","generated_at":"2026-09-01T20:11:18.457385+00:00","sources":["arxiv_cs_cl","arxiv_llm_reliability"],"days":[{"date":"2026-08-17","items":[{"title":"FTA-Mem: Fact-Time-Affect Anchored Memory for Low-Density Long-Term Dialogue","url":"http://arxiv.org/abs/2608.16303v1","source":"arxiv_llm_reliability","type":"paper","summary_1line":"Long-term emotional-support agents require memory mechanisms for personalized understanding across sessions. However, emotional-support dialogue is often low-density: turns are incomplete, evidence is scattered, and u...","why_it_matters":"Matches feed focus: agent, eval.","sid":"40602cd71e370eb6","published":"2026-08-17T09:14:31+00:00","editor_note":"Opens the thread with a memory architecture purpose-built for low-density emotional-support dialogue, not general-purpose recall."}]},{"date":"2026-08-30","items":[{"title":"Agent Zero Memory: Provenance-Aware Long-Term Memory for LLM Agents","url":"http://arxiv.org/abs/2608.29606v1","source":"arxiv_cs_cl","type":"paper","summary_1line":"Large language model (LLM) agents need durable, faithful memory of everything a user or organization has said and stored, yet most memory systems commit to a single organizing structure (a fact store, a vector index,...","why_it_matters":"Matches feed focus: agentic, eval.","sid":"5960e24b491051f2","published":"2026-08-30T06:55:59+00:00","editor_note":"Argues against committing to one memory structure, proposing provenance tracking across fact stores, vector indexes, and other backends."}]},{"date":"2026-08-31","items":[{"title":"UTILMEM: Benchmarking Evidence Utilization in Long-Term Conversational Memory","url":"http://arxiv.org/abs/2608.30508v1","source":"arxiv_llm_reliability","type":"paper","summary_1line":"Long-term memory is increasingly important for conversational agents, yet existing benchmarks primarily measure memory through pointwise factual recall: whether a system can recover isolated facts or event-level detai...","why_it_matters":"Matches feed focus: agent, eval.","sid":"92250613f04ac1b9","published":"2026-08-31T09:41:26+00:00","editor_note":"Reframes evaluation itself: scoring whether an agent actually uses retrieved evidence, not just whether it recalls isolated facts."}]}],"editorial":{"tldr":"Three separate long-term-memory research papers grouped by a shared title phrase rather than one developing story. It opened Aug 17 with FTA-Mem's fact-time-affect anchored memory for low-density emotional-support dialogue, then Agent Zero Memory's Aug 30 case for provenance-aware memory instead of one fixed structure.","stale":false,"whats_new":"A new arXiv benchmark, UTILMEM, argues existing long-term-memory evals measure only pointwise factual recall and proposes scoring how well an agent actually uses the evidence it retrieves.","why_it_matters":"If you eval your agent's memory system on recall alone, UTILMEM's framing is a reminder that retrieving the right fact and correctly using it in the reply are different failure modes worth separate checks.","take_for_builders":"Before trusting a recall-only memory eval, check whether it also scores evidence utilization — UTILMEM's task design is a starting point for adding that check to your own harness.","open_questions":["Does FTA-Mem's fact-time-affect anchoring generalize beyond emotional-support dialogue to task-oriented agents?","Does Agent Zero Memory's provenance tracking add meaningful latency or storage overhead versus a single fixed structure?","Does UTILMEM report results for today's popular memory frameworks (vector-store or graph-based), or only research prototypes?"],"generated_at":"2026-09-01T05:10:00Z"}}