{
  "schema_version": 1,
  "builder_plan": {
    "scenario": "A support or data assistant repeats work across related customer requests.",
    "change": "Try the paper's memory change in a replay of past tasks, with separate memory per customer.",
    "baseline": "Your current memory pipeline plus a no-memory control using the same agent.",
    "steps": [
      "Locate the paper's implementation and check its license, model support, data needs, and hardware requirements. Stop here if you cannot obtain the required artifacts.",
      "Save a small set of representative product tasks and expected outcomes. Keep evaluation tasks out of any training or tuning, and choose quality and cost limits before running.",
      "Run the baseline and the proposed change on identical inputs. Keep other settings fixed and repeat paired runs to expose variation.",
      "Save per-task outputs, failures, timings, and all usage costs. Review regressions before expanding the test."
    ],
    "measure": "Task success, stale memories, retrieval misses, response latency, and all memory maintenance costs.",
    "decision": "Keep the change only if it meets your preselected quality and latency limits and improves the metric you care about after setup and operating costs. Otherwise retain the baseline. This is an exploration plan, not a validated recommendation."
  },
  "builder_review": null,
  "arxiv_id": "2610.29671",
  "title": "Uni-Mem: Learning a Long-Horizon Agentic Memory Manager with Unified Online Policy Optimization",
  "summary": "Replaces hand-designed memory workflows and hybrid RL-plus-offline-supervision pipelines with a single policy trained end-to-end online for both memory construction and retrieval, using stage-decoupled group comparisons and event-level credit assignment. Costs include an online RL training loop over long memory trajectories, the added complexity of alternating stage optimization, and dependence on RL-quality reward signals; the abstract gives no compute or latency figures. Can fail if rewards are sparse or misaligned, if benchmarks like LoCoMo do not transfer to the reader's workloads, or if the trained policy is brittle under distribution shift in production interactions.",
  "rationale": "The method requires online RL infrastructure and reward design beyond a 24 GB GPU or API-only setup, and the paper reports no efficiency gains or concrete improvement factors that would justify adoption or small-scale testing now.",
  "technique": "Stage-decoupled RL for unified agent memory management",
  "family": "memoria_e_contexto",
  "recommendation": "observar",
  "published": "2026-10-21",
  "as_of": "2026-10-22",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2610.29671",
  "independent_implementations": 0,
  "total_implementations": 0,
  "stars": 0,
  "citations": null,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "Replaces hand-designed memory workflows and hybrid RL-plus-offline-supervision pipelines with a single policy trained end-to-end online for both memory construction and retrieval, using stage-decoupled group comparisons and event-level credit assignment. Costs include an online RL training loop over long memory trajectories, the added complexity of alternating stage optimization, and dependence on RL-quality reward signals; the abstract gives no compute or latency figures. Can fail if rewards are sparse or misaligned, if benchmarks like LoCoMo do not transfer to the reader's workloads, or if the trained policy is brittle under distribution shift in production interactions.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.29671",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "claim_id": "claim-02",
      "statement": "Uni-Mem outperforms the strongest RL-based memory baselines on LoCoMo and LongMemEval; no quantitative factor is reported in the abstract.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.29671",
      "source_page": null,
      "source_excerpt": ""
    }
  ],
  "exposure_map": [
    {
      "dimension": "quality",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "compute",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "latency",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "operations",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "compatibility",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "security",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "data_and_training",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "dimension": "reproducibility",
      "basis": "not_evaluated",
      "finding": "",
      "interpretation": "",
      "limitation": "",
      "next_check": "",
      "source_url": "",
      "source_page": null,
      "source_excerpt": ""
    }
  ],
  "risks": [],
  "minimum_test": [],
  "open_questions": [],
  "independent_tests": [],
  "lab_reproductions": [],
  "report_available": false,
  "validation_tier": "unknown",
  "evidence_tier": "unknown",
  "infrastructure_basis": "unknown",
  "equations": [],
  "equations_status": "not_fetched",
  "core_kind": "",
  "equations_fetched_at": "",
  "equations_source_url": "",
  "equations_source_sha256": "",
  "exposures_reviewed_at": "",
  "exposures_source_sha256": ""
}