{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:3PYGNUK7ETWFJSTA2S64NJZQWA","short_pith_number":"pith:3PYGNUK7","schema_version":"1.0","canonical_sha256":"dbf066d15f24ec54ca60d4bdc6a730b034b3cccb910d09803e1a02354903bbee","source":{"kind":"arxiv","id":"2107.08353","version":4},"attestation_state":"computed","paper":{"title":"Top-label calibration and multiclass-to-binary reductions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ME","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aaditya Ramdas, Chirag Gupta","submitted_at":"2021-07-18T03:27:50Z","abstract_excerpt":"A multiclass classifier is said to be top-label calibrated if the reported probability for the predicted class -- the top-label -- is calibrated, conditioned on the top-label. This conditioning on the top-label is absent in the closely related and popular notion of confidence calibration, which we argue makes confidence calibration difficult to interpret for decision-making. We propose top-label calibration as a rectification of confidence calibration. Further, we outline a multiclass-to-binary (M2B) reduction framework that unifies confidence, top-label, and class-wise calibration, among othe"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2107.08353","kind":"arxiv","version":4},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-07-18T03:27:50Z","cross_cats_sorted":["cs.AI","stat.ME","stat.ML"],"title_canon_sha256":"5136b5bd717fa03a970a7066d4e63eff0b57aec596e17b75fd83c64a522a863d","abstract_canon_sha256":"ffa9feff2e70fb0ac3e26fd96b6ce62d8aa8712663ee2d0403cf52fab1bceae1"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:54:15.090251Z","signature_b64":"TqEicFcIrcvlKEjLt7svuoPET9pxvPwFJ4Zl8cu3RXOSaH5DtZXVg5dQ682Y2TNF8VCLHntWj39UklVukULCBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"dbf066d15f24ec54ca60d4bdc6a730b034b3cccb910d09803e1a02354903bbee","last_reissued_at":"2026-07-05T04:54:15.089840Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:54:15.089840Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Top-label calibration and multiclass-to-binary reductions","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ME","stat.ML"],"primary_cat":"cs.LG","authors_text":"Aaditya Ramdas, Chirag Gupta","submitted_at":"2021-07-18T03:27:50Z","abstract_excerpt":"A multiclass classifier is said to be top-label calibrated if the reported probability for the predicted class -- the top-label -- is calibrated, conditioned on the top-label. This conditioning on the top-label is absent in the closely related and popular notion of confidence calibration, which we argue makes confidence calibration difficult to interpret for decision-making. We propose top-label calibration as a rectification of confidence calibration. Further, we outline a multiclass-to-binary (M2B) reduction framework that unifies confidence, top-label, and class-wise calibration, among othe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2107.08353","kind":"arxiv","version":4},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2107.08353/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2107.08353","created_at":"2026-07-05T04:54:15.089894+00:00"},{"alias_kind":"arxiv_version","alias_value":"2107.08353v4","created_at":"2026-07-05T04:54:15.089894+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2107.08353","created_at":"2026-07-05T04:54:15.089894+00:00"},{"alias_kind":"pith_short_12","alias_value":"3PYGNUK7ETWF","created_at":"2026-07-05T04:54:15.089894+00:00"},{"alias_kind":"pith_short_16","alias_value":"3PYGNUK7ETWFJSTA","created_at":"2026-07-05T04:54:15.089894+00:00"},{"alias_kind":"pith_short_8","alias_value":"3PYGNUK7","created_at":"2026-07-05T04:54:15.089894+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.17200","citing_title":"Calibrating Model-Based Evaluation Metrics for Summarization","ref_index":141,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/3PYGNUK7ETWFJSTA2S64NJZQWA","json":"https://pith.science/pith/3PYGNUK7ETWFJSTA2S64NJZQWA.json","graph_json":"https://pith.science/api/pith-number/3PYGNUK7ETWFJSTA2S64NJZQWA/graph.json","events_json":"https://pith.science/api/pith-number/3PYGNUK7ETWFJSTA2S64NJZQWA/events.json","paper":"https://pith.science/paper/3PYGNUK7"},"agent_actions":{"view_html":"https://pith.science/pith/3PYGNUK7ETWFJSTA2S64NJZQWA","download_json":"https://pith.science/pith/3PYGNUK7ETWFJSTA2S64NJZQWA.json","view_paper":"https://pith.science/paper/3PYGNUK7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2107.08353&json=true","fetch_graph":"https://pith.science/api/pith-number/3PYGNUK7ETWFJSTA2S64NJZQWA/graph.json","fetch_events":"https://pith.science/api/pith-number/3PYGNUK7ETWFJSTA2S64NJZQWA/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/3PYGNUK7ETWFJSTA2S64NJZQWA/action/timestamp_anchor","attest_storage":"https://pith.science/pith/3PYGNUK7ETWFJSTA2S64NJZQWA/action/storage_attestation","attest_author":"https://pith.science/pith/3PYGNUK7ETWFJSTA2S64NJZQWA/action/author_attestation","sign_citation":"https://pith.science/pith/3PYGNUK7ETWFJSTA2S64NJZQWA/action/citation_signature","submit_replication":"https://pith.science/pith/3PYGNUK7ETWFJSTA2S64NJZQWA/action/replication_record"}},"created_at":"2026-07-05T04:54:15.089894+00:00","updated_at":"2026-07-05T04:54:15.089894+00:00"}