{"record": {"aliases": ["CalArena: A Large-Scale Post-Hoc Calibration Benchmark"], "artifacts": [{"id": "arxiv:2605.30188", "kind": "paper", "url": "https://arxiv.org/abs/2605.30188"}], "authors": ["Eugène Berta", "David Holzmüller", "Francis Bach", "Michael I. Jordan"], "authors_reference": {"basis": "arxiv_metadata", "retrieved_at": "2026-11-03", "source_key": "claire-radar:2605.30188", "source_url": "https://arxiv.org/abs/2605.30188"}, "categories": ["General AI", "Vision & 3D", "cs.LG"], "description": {"en": "CalArena is a standardized benchmark for comparing post-hoc calibration methods across tabular and computer-vision classification settings."}, "documents": [{"document_type": "registry_page", "id": "claire_radar:https://arxiv.org/abs/2605.30188", "source": "claire_radar", "source_url": "https://arxiv.org/abs/2605.30188", "title": "CalArena"}], "evidence_summary": {"document_count": 1, "model_count": null, "model_count_basis": null}, "institutions": ["Institut national de recherche en sciences et technologies du numérique"], "institutions_reference": {"basis": "openalex_linked_institutions", "doi": "10.48550/arxiv.2605.30188", "retrieved_at": "2026-11-04", "source_key": "claire-radar:2605.30188", "source_url": "https://openalex.org/W7162784328"}, "key": "claire-radar:2605.30188", "modality": null, "name": "CalArena", "openness": {"code_license": null, "data_license": null, "evidence": [], "status": "unknown"}, "provenance": {"admission_policy_version": "2026-09-09.1", "confidence": "Medium", "crawl_bundle": "claire_radar_2026-10-14", "crawled_at": "2026-09-25T04:40:22Z", "data_status": "primary-source-indexed", "display_eligible": "true", "origin_record_id": "bm_calarena-a-large-scale-post-hoc-calibratio_c7e8f107", "origin_source": "arxiv", "recognition_confidence": "0.6", "record_sha256": "07d175fef252350aa2efe64cc3eedb370a751cf39cd059416d8f35a1f4b2dc9f", "relation": "introduces", "review_model": "codex-cli-default", "review_state": "ai-reviewed", "reviewed_at": "2026-09-24T18:07:12.485829Z", "source_url": "https://arxiv.org/abs/2605.30188"}, "publisher": null, "released": "2026-05-28", "released_reference": {"basis": "benchmark_release", "source_key": "claire-radar:2605.30188", "source_url": "https://arxiv.org/abs/2605.30188"}, "schema_version": 1, "sizes": [], "slug": "claire-radar-2605-30188", "source": "claire_radar", "source_benchmark_id": "2605.30188", "source_metadata": {"claire_radar": {"adoption30d": null, "aliases": ["CalArena: A Large-Scale Post-Hoc Calibration Benchmark"], "annotation": "Mixed", "applicationDomains": ["General AI"], "area": "Vision & 3D", "attentionForecast": {"confidence": "Medium", "horizon": "7d", "reason": "Its broad experimental scope, explicit metric, and plug-and-play release make it readily usable by calibration researchers.", "score": 67}, "capabilities": [], "confidence": "Medium", "construction": "Aggregate Existing", "constructionDetail": "Unknown — the indexer does not infer construction details without explicit source evidence.", "copyGeneration": {"generatedAt": "2026-09-24T18:07:12.485829Z", "inputHash": "5684b64c78de53e9304d7468d3450630cf8876f1b68936644393c19ade55d7b9", "model": "codex-cli-default"}, "curation": {"admissionPolicyVersion": "2026-09-09.1", "canonicalNameEvidence": "CalArena: A Large-Scale Post-Hoc Calibration Benchmark", "canonicalNameSource": "paper_title", "decisionReason": "The supplied text provides a formal benchmark name, a defined evaluation task, an explicit scoring contract, and a public reusable release.", "evaluationEvidence": {"reuse": "we release all data, code, and evaluation tools, providing a plug-and-play benchmark for developing and comparing calibration methods", "scoring": "Post-Hoc Improvement (PHI) in proper scoring rules", "task": "a large-scale, standardized benchmark for post-hoc calibration"}, "model": "codex-cli-default", "reviewedAt": "2026-09-24T18:07:12.485829Z", "state": "ai-reviewed"}, "dataStatus": "primary-source-indexed", "demo": false, "description": "CalArena is a standardized benchmark for comparing post-hoc calibration methods across tabular and computer-vision classification settings.", "displayEligible": true, "domainCuration": {"method": "rules-v1", "state": "auto"}, "evaluationMode": "public_reusable", "evidence": {"reasonCodes": ["benchmark term in abstract", "benchmark release stated in one sentence", "evaluation protocol evidence"], "snippet": "We introduce a large-scale, standardized benchmark for post-hoc calibration, covering nearly 2000 experiments across tabular and computer vision tasks, including binary, multiclass, and large-scale classification settings."}, "familyId": "bmf_6cdd55a37093", "firstSeenAt": "2026-08-19", "heat": null, "id": "bm_calarena-a-large-scale-post-hoc-calibratio_c7e8f107", "indexedAt": "2026-08-19T11:05:13.395391Z", "industrySectors": [], "links": {"code": null, "data": null, "paper": "https://arxiv.org/abs/2605.30188", "pdf": "https://arxiv.org/pdf/2605.30188", "project": null, "report": "https://arxiv.org/abs/2605.30188"}, "metrics": [], "motivation": "Reliable probability estimates are critical in many machine learning applications, yet modern classifiers are often poorly calibrated.", "name": "CalArena", "oneLine": "CalArena is a standardized benchmark for comparing post-hoc calibration methods across tabular and computer-vision classification settings.", "paperTitle": "CalArena: A Large-Scale Post-Hoc Calibration Benchmark", "primaryDomain": "General AI", "publication": {"evidence": null, "evidenceLevel": "unverified", "evidenceUrl": "https://arxiv.org/abs/2605.30188", "source": "arxiv-metadata", "status": "unverified", "venue": null, "verifiedAt": "2026-08-19T11:05:13.395391Z"}, "publications": [], "readiness": "Paper only", "recognitionConfidence": 0.6, "relation": "introduces", "releasedAt": "2026-05-28", "reviewLifecycle": {"state": "pending"}, "source": {"authors": ["Eug\\`ene Berta", "David Holzm\\\"uller", "Francis Bach", "Michael I. Jordan"], "categories": ["cs.LG", "cs.AI", "stat.ML"], "id": "2605.30188", "title": "CalArena: A Large-Scale Post-Hoc Calibration Benchmark", "type": "arxiv", "url": "https://arxiv.org/abs/2605.30188"}, "sourceUpdatedAt": "2026-06-01T16:20:48Z", "topics": ["cs.LG"], "venueAttempts": [], "whyItMatters": "It supplies a common evaluation framework and a principled scoring measure for comparing calibration methods at scale."}}}, "schema_version": 1, "scores_by_source": {}, "siblings": []}
