{ "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": "2610.23598", "title": "RL-Guided Temporal Localization for Dual-Channel Retrieval in Long-Horizon Agent Memory", "summary": "TIDER replaces purely lexical/semantic retrieval over agent long-term memory (and approaches that inject time only at organization or generation) with an RL-trained temporal interval locator whose predicted windows define a temporal retrieval channel, merged with semantic results and reranked by temporal proximity. The cost is training a 1.7B locator with composite RL rewards (feasible on a 24 GB GPU but non-trivial engineering), plus added retrieval latency and pipeline complexity from the dual-channel merge and reranking. It can fail when queries lack identifiable temporal structure, when timestamps in memory are noisy or missing, or when the locator mispredicts windows and excludes the correct evidence before the semantic channel can compensate; benchmark gains on LifeBench/LoCoMo may not transfer to other memory distributions.", "rationale": "The accuracy gains are meaningful and the 1.7B locator fits constrained infrastructure, but the RL training pipeline and dependence on timestamped, temporally structured memories require validation on the reader's own workload before production adoption.", "technique": "RL-trained temporal window localization for dual-channel memory retrieval", "family": "memoria_e_contexto", "recommendation": "testar", "published": "2026-10-16", "as_of": "2026-10-19", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.23598", "independent_implementations": 2, "total_implementations": 3, "stars": 132, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "TIDER replaces purely lexical/semantic retrieval over agent long-term memory (and approaches that inject time only at organization or generation) with an RL-trained temporal interval locator whose predicted windows define a temporal retrieval channel, merged with semantic results and reranked by temporal proximity. The cost is training a 1.7B locator with composite RL rewards (feasible on a 24 GB GPU but non-trivial engineering), plus added retrieval latency and pipeline complexity from the dual-channel merge and reranking. It can fail when queries lack identifiable temporal structure, when timestamps in memory are noisy or missing, or when the locator mispredicts windows and excludes the correct evidence before the semantic channel can compensate; benchmark gains on LifeBench/LoCoMo may not transfer to other memory distributions.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.23598", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "Reports state-of-the-art accuracy on two long-horizon memory benchmarks: 61.52% on LifeBench (+8.82 points) and 78.96% on LoCoMo (+3.66 points) using a 1.7B temporal locator.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.23598", "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": "" }