{"paper":{"title":"Task-Agnostic Noisy Label Detection via Standardized Loss Aggregation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Aggregating standardized cross-validation losses produces reliable sample-level scores for detecting noisy labels.","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Doohyun Park, Inhyuk Park","submitted_at":"2026-05-11T08:16:09Z","abstract_excerpt":"Noisy labels are common in large-scale medical imaging datasets due to inter-observer variability and ambiguous cases. We propose a statistically grounded and task-agnostic framework, Standardized Loss Aggregation (SLA), for detecting noisy labels at the sample level. SLA quantifies label reliability by aggregating standardized fold-level validation losses across repeated cross-validation runs. This formulation generalizes discrete hard-counting schemes into a continuous estimator that captures both the frequency and magnitude of performance deviations, yielding interpretable and statistically"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"SLA generalizes discrete hard-counting schemes into a continuous estimator that captures both the frequency and magnitude of performance deviations, yielding interpretable and statistically stable noisiness scores.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That deviations in fold-level validation losses across cross-validation runs are caused primarily by label noise rather than by model randomness, data split effects, or other unmodeled factors.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SLA converts hard-counting of high-loss samples into a continuous noisiness score by standardizing fold-level validation losses and aggregating them over multiple cross-validation runs, showing better performance than baselines on fundus data.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Aggregating standardized cross-validation losses produces reliable sample-level scores for detecting noisy labels.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"6039bdddbc4af920c394f01f6b74388cdf76580b00ad62685e648e6137a3cb3c"},"source":{"id":"2605.10165","kind":"arxiv","version":2},"verdict":{"id":"7e2fe95e-f73b-4411-bd62-4e6390068d5b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-12T03:03:25.559556Z","strongest_claim":"SLA generalizes discrete hard-counting schemes into a continuous estimator that captures both the frequency and magnitude of performance deviations, yielding interpretable and statistically stable noisiness scores.","one_line_summary":"SLA converts hard-counting of high-loss samples into a continuous noisiness score by standardizing fold-level validation losses and aggregating them over multiple cross-validation runs, showing better performance than baselines on fundus data.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That deviations in fold-level validation losses across cross-validation runs are caused primarily by label noise rather than by model randomness, data split effects, or other unmodeled factors.","pith_extraction_headline":"Aggregating standardized cross-validation losses produces reliable sample-level scores for detecting noisy labels."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.10165/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"claim_evidence","ran_at":"2026-05-20T06:22:00.948767Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T15:39:14.066848Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T12:01:17.495592Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T09:38:21.616054Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"df4f8794a11a834459c471225f4ca0ef5c4cdf7635385944d1e034d2639775d5"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}