{
  "schema_version": 1,
  "builder_plan": {
    "scenario": "A narrow, repeated product task may not need your largest model.",
    "change": "Check whether a released student model fits the task before generating training data.",
    "baseline": "Your current model and an off-the-shelf small model on the same held-out tasks.",
    "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": "Student quality, serving cost, teacher calls, training cost, and failures on unusual inputs.",
    "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": "2609.27307",
  "title": "Learn How to Act from Your Own Interactions: On-Policy Self-Distillation for GUI Agents",
  "summary": "GUI-SD-v2 replaces external teacher or human-annotated supervision for multi-turn GUI agents with a two-stage on-policy self-distillation pipeline: joint rollout optimization with and without privileged guidance, followed by selective distillation of step-specific reasoning and memory from a privilege-conditioned self-teacher. The cost is a full RL-style training loop with environment rollouts on AndroidWorld/MobileWorld-class simulators, two training stages, and trajectory data generation, which exceeds a single 24 GB GPU budget for any non-trivial base model and adds substantial engineering complexity. Failure modes include the self-teacher's imperfect privilege following propagating errors into the student, overfitting to the two evaluated mobile benchmarks, and memory-guidance distillation degrading on long-horizon tasks outside the training distribution.",
  "rationale": "The method is a multi-stage training recipe requiring on-policy rollouts and simulator infrastructure that does not fit a 24 GB, no-cluster budget, so it is worth tracking for its released code and data rather than adopting or testing now.",
  "technique": "On-policy self-distillation for multi-turn GUI agents",
  "family": "destilacao",
  "recommendation": "observar",
  "published": "2026-09-23",
  "as_of": "2026-10-18",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2609.27307",
  "independent_implementations": 0,
  "total_implementations": 1,
  "stars": 874,
  "citations": 0,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "GUI-SD-v2 replaces external teacher or human-annotated supervision for multi-turn GUI agents with a two-stage on-policy self-distillation pipeline: joint rollout optimization with and without privileged guidance, followed by selective distillation of step-specific reasoning and memory from a privilege-conditioned self-teacher. The cost is a full RL-style training loop with environment rollouts on AndroidWorld/MobileWorld-class simulators, two training stages, and trajectory data generation, which exceeds a single 24 GB GPU budget for any non-trivial base model and adds substantial engineering complexity. Failure modes include the self-teacher's imperfect privilege following propagating errors into the student, overfitting to the two evaluated mobile benchmarks, and memory-guidance distillation degrading on long-horizon tasks outside the training distribution.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2609.27307",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "claim_id": "claim-02",
      "statement": "Outperforms evaluated state-of-the-art methods in Pass@1 and Pass@3 success rates on AndroidWorld and MobileWorld; no quantitative factor reported in the abstract.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2609.27307",
      "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": ""
}