{
  "schema_version": 1,
  "builder_plan": {
    "scenario": "Start with one repeated product task that has a measurable failure or cost.",
    "change": "Read the methods and identify one change you can isolate in that workflow.",
    "baseline": "Your current implementation 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": "Successful outcomes, response time, human rework, and total cost per successful task.",
    "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": "2610.09802",
  "title": "DisParQ: Self-Supervised Part Concepts for Interpretable Vision Foundation Models",
  "summary": "DisParQ replaces opaque patch embeddings or language-dependent concept bottlenecks with a learned discrete prototype dictionary plus quantized attribute residuals trained to reconstruct a frozen self-supervised backbone's features, without class labels or text supervision. The cost is an additional training stage (prototype dictionary, sparse assignment, spatial reconstruction decoder) on top of the frozen backbone, plus inference overhead from patch-to-concept assignment and the constraint that the concept layer may discard information the backbone carried. It can fail if the learned concepts do not align with parts humans care about on a new domain, if the sparsity constraint under-represents atypical images, or if reconstruction fidelity does not translate into faithful explanations of downstream predictions.",
  "rationale": "The method is feasible on constrained infrastructure since the backbone stays frozen, but its value is interpretability rather than an efficiency or accuracy gain, so it only warrants adoption when auditable concept-level explanations are an explicit production requirement.",
  "technique": "Discrete part-based interpretable concept layer on frozen ViT",
  "family": "outro",
  "recommendation": "observar",
  "published": "2026-10-07",
  "as_of": "2026-10-18",
  "editorial_status": "indexed",
  "source_url": "https://arxiv.org/abs/2610.09802",
  "independent_implementations": 0,
  "total_implementations": 1,
  "stars": 121,
  "citations": null,
  "claims": [
    {
      "claim_id": "claim-01",
      "statement": "DisParQ replaces opaque patch embeddings or language-dependent concept bottlenecks with a learned discrete prototype dictionary plus quantized attribute residuals trained to reconstruct a frozen self-supervised backbone's features, without class labels or text supervision. The cost is an additional training stage (prototype dictionary, sparse assignment, spatial reconstruction decoder) on top of the frozen backbone, plus inference overhead from patch-to-concept assignment and the constraint that the concept layer may discard information the backbone carried. It can fail if the learned concepts do not align with parts humans care about on a new domain, if the sparsity constraint under-represents atypical images, or if reconstruction fidelity does not translate into faithful explanations of downstream predictions.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.09802",
      "source_page": null,
      "source_excerpt": ""
    },
    {
      "claim_id": "claim-02",
      "statement": "Matches its frozen DINOv2 teacher on ImageNet linear probing at 83.2% top-1, i.e., parity with the backbone while adding discrete interpretable concepts; higher concept consistency than language-aligned models is claimed but not given as a quantified factor.",
      "basis": "inferred",
      "result": "",
      "baseline": "",
      "conditions": "",
      "source_url": "https://arxiv.org/abs/2610.09802",
      "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": ""
}