{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:4T4CYBYPIUTDDHTZR6KOTEK7IM","short_pith_number":"pith:4T4CYBYP","schema_version":"1.0","canonical_sha256":"e4f82c070f4526319e798f94e9915f43061a6c2bb9482435d3112d304a9cf28b","source":{"kind":"arxiv","id":"2210.16955","version":2},"attestation_state":"computed","paper":{"title":"Learning to Defer to Multiple Experts: Consistent Surrogate Losses, Confidence Calibration, and Conformal Ensembles","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Daniel Barrej\\'on, Eric Nalisnick, Rajeev Verma","submitted_at":"2022-10-30T21:27:29Z","abstract_excerpt":"We study the statistical properties of learning to defer (L2D) to multiple experts. In particular, we address the open problems of deriving a consistent surrogate loss, confidence calibration, and principled ensembling of experts. Firstly, we derive two consistent surrogates -- one based on a softmax parameterization, the other on a one-vs-all (OvA) parameterization -- that are analogous to the single expert losses proposed by Mozannar and Sontag (2020) and Verma and Nalisnick (2022), respectively. We then study the frameworks' ability to estimate P( m_j = y | x ), the probability that the jth"},"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":"2210.16955","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2022-10-30T21:27:29Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"28048c284dea4a9b2a133d98eb778224cefe9de459b9957ce6de82c2a56ee3b1","abstract_canon_sha256":"4e0b08a75af1ae22d90daf79bb2c1474471b10bf31a50f4722d4a27fb8a8341f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:44:51.184192Z","signature_b64":"4l/nxWQ125MviW9Omuht/qLb6bP5FbRxFM2I760l903ZcWyX8y2dhwW3klqVAbu5QSdjnjyUWcF2cKzG8qc4BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e4f82c070f4526319e798f94e9915f43061a6c2bb9482435d3112d304a9cf28b","last_reissued_at":"2026-07-05T05:44:51.183661Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:44:51.183661Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Learning to Defer to Multiple Experts: Consistent Surrogate Losses, Confidence Calibration, and Conformal Ensembles","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Daniel Barrej\\'on, Eric Nalisnick, Rajeev Verma","submitted_at":"2022-10-30T21:27:29Z","abstract_excerpt":"We study the statistical properties of learning to defer (L2D) to multiple experts. In particular, we address the open problems of deriving a consistent surrogate loss, confidence calibration, and principled ensembling of experts. Firstly, we derive two consistent surrogates -- one based on a softmax parameterization, the other on a one-vs-all (OvA) parameterization -- that are analogous to the single expert losses proposed by Mozannar and Sontag (2020) and Verma and Nalisnick (2022), respectively. We then study the frameworks' ability to estimate P( m_j = y | x ), the probability that the jth"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.16955","kind":"arxiv","version":2},"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/2210.16955/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":"2210.16955","created_at":"2026-07-05T05:44:51.183718+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.16955v2","created_at":"2026-07-05T05:44:51.183718+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.16955","created_at":"2026-07-05T05:44:51.183718+00:00"},{"alias_kind":"pith_short_12","alias_value":"4T4CYBYPIUTD","created_at":"2026-07-05T05:44:51.183718+00:00"},{"alias_kind":"pith_short_16","alias_value":"4T4CYBYPIUTDDHTZ","created_at":"2026-07-05T05:44:51.183718+00:00"},{"alias_kind":"pith_short_8","alias_value":"4T4CYBYP","created_at":"2026-07-05T05:44:51.183718+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.12587","citing_title":"Strategic Decision Support for AI Agents","ref_index":89,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4T4CYBYPIUTDDHTZR6KOTEK7IM","json":"https://pith.science/pith/4T4CYBYPIUTDDHTZR6KOTEK7IM.json","graph_json":"https://pith.science/api/pith-number/4T4CYBYPIUTDDHTZR6KOTEK7IM/graph.json","events_json":"https://pith.science/api/pith-number/4T4CYBYPIUTDDHTZR6KOTEK7IM/events.json","paper":"https://pith.science/paper/4T4CYBYP"},"agent_actions":{"view_html":"https://pith.science/pith/4T4CYBYPIUTDDHTZR6KOTEK7IM","download_json":"https://pith.science/pith/4T4CYBYPIUTDDHTZR6KOTEK7IM.json","view_paper":"https://pith.science/paper/4T4CYBYP","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.16955&json=true","fetch_graph":"https://pith.science/api/pith-number/4T4CYBYPIUTDDHTZR6KOTEK7IM/graph.json","fetch_events":"https://pith.science/api/pith-number/4T4CYBYPIUTDDHTZR6KOTEK7IM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4T4CYBYPIUTDDHTZR6KOTEK7IM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4T4CYBYPIUTDDHTZR6KOTEK7IM/action/storage_attestation","attest_author":"https://pith.science/pith/4T4CYBYPIUTDDHTZR6KOTEK7IM/action/author_attestation","sign_citation":"https://pith.science/pith/4T4CYBYPIUTDDHTZR6KOTEK7IM/action/citation_signature","submit_replication":"https://pith.science/pith/4T4CYBYPIUTDDHTZR6KOTEK7IM/action/replication_record"}},"created_at":"2026-07-05T05:44:51.183718+00:00","updated_at":"2026-07-05T05:44:51.183718+00:00"}