{
  "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": "2609.35532",
  "title": "ARISE: Adapting to Evolving Capability Gaps in Agentic Reinforcement Learning",
  "summary": "ARISE replaces fixed evaluation criteria and static task sampling in agentic reinforcement learning with co-evolving rubrics, refined exploration skills, and capability-adaptive task sampling driven by rollout evidence. It costs the full RL training loop itself—rollout generation, rubric and skill updates, and policy optimization—which is compute far beyond a single 24 GB GPU, plus substantial implementation complexity in the evolving evaluation machinery. It can fail if evolved rubrics drift toward rewarding partial behaviors that do not correlate with real task success, or if adaptive sampling over-focuses on current weaknesses and starves coverage of previously mastered capabilities.",
  "rationale": "The method targets teams already running RL training for long-horizon agents, which the reader's single-GPU, no-foundation-model-training constraint cannot support, so it is worth tracking rather than adopting.",
  "technique": "Adaptive rubric-skill co-evolution for agentic RL",
  "family": "treino_eficiente",
  "recommendation": "observar",
  "published": "2026-09-28",
  "as_of": "2026-10-18",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2609.35532",
  "independent_implementations": 0,
  "total_implementations": 1,
  "stars": 121,
  "citations": null,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "ARISE replaces fixed evaluation criteria and static task sampling in agentic reinforcement learning with co-evolving rubrics, refined exploration skills, and capability-adaptive task sampling driven by rollout evidence. It costs the full RL training loop itself—rollout generation, rubric and skill updates, and policy optimization—which is compute far beyond a single 24 GB GPU, plus substantial implementation complexity in the evolving evaluation machinery. It can fail if evolved rubrics drift toward rewarding partial behaviors that do not correlate with real task success, or if adaptive sampling over-focuses on current weaknesses and starves coverage of previously mastered capabilities.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2609.35532",
      "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": ""
}