{ "schema_version": 1, "builder_plan": { "scenario": "You already need task-specific training and want to reduce its memory or compute cost.", "change": "Change only the training technique on a small controlled run after checking prerequisites.", "baseline": "Your current training recipe with the same model, data split, and evaluation.", "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": "Held-out quality, peak memory, training time, and full compute cost to the target quality.", "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": "2610.22344", "title": "MLExpert: Towards a Generalist Agentic Model for Autonomous Machine Learning Engineering", "summary": "The method replaces workflow-scripted MLE orchestration and full-trajectory RL with hierarchical agentic SFT plus windowed RL that assigns shared outcome rewards to decision windows between experiment results. It costs a dedicated task factory (Kaggle-MLE Factory), executable interaction harness, asynchronous rollouts for a 35B MoE model, and substantial RL compute — all beyond a single 24 GB GPU or API budget. If adopted indirectly (e.g., via released weights or scaffold ideas), failure modes include reward misattribution across long windows, overfitting to Kaggle-style tasks, and costly runaway experimentation loops.", "rationale": "The technique requires foundation-scale post-training infrastructure unavailable to the reader, so its value is limited to monitoring for released checkpoints or adopting its evaluation benchmarks.", "technique": "Windowed Interaction RL for MLE agent post-training", "family": "treino_eficiente", "recommendation": "observar", "published": "2026-10-15", "as_of": "2026-10-18", "editorial_status": "indexed", "source_url": "https://arxiv.org/abs/2610.22344", "independent_implementations": 2, "total_implementations": 3, "stars": 132, "citations": null, "claims": [ { "claim_id": "claim-01", "statement": "The method replaces workflow-scripted MLE orchestration and full-trajectory RL with hierarchical agentic SFT plus windowed RL that assigns shared outcome rewards to decision windows between experiment results. It costs a dedicated task factory (Kaggle-MLE Factory), executable interaction harness, asynchronous rollouts for a 35B MoE model, and substantial RL compute — all beyond a single 24 GB GPU or API budget. If adopted indirectly (e.g., via released weights or scaffold ideas), failure modes include reward misattribution across long windows, overfitting to Kaggle-style tasks, and costly runaway experimentation loops.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.22344", "source_page": null, "source_excerpt": "" }, { "claim_id": "claim-02", "statement": "On MLE-Bench Lite, MLExpert raises the base model's average medal rate from 37.9% to 72.7% (a 1.92x ratio), reported as comparable to larger frontier models.", "basis": "inferred", "result": "", "baseline": "", "conditions": "", "source_url": "https://arxiv.org/abs/2610.22344", "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": "" }