{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:IQXZYQE7WCK2T64EWBT5I6HYGF","short_pith_number":"pith:IQXZYQE7","schema_version":"1.0","canonical_sha256":"442f9c409fb095a9fb84b067d478f83151c835afc36d5e19aa1618b9748e52ad","source":{"kind":"arxiv","id":"2303.02970","version":1},"attestation_state":"computed","paper":{"title":"Rethinking Confidence Calibration for Failure Prediction","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Cheng-Lin Liu, Fei Zhu, Xu-Yao Zhang, Zhen Cheng","submitted_at":"2023-03-06T08:54:18Z","abstract_excerpt":"Reliable confidence estimation for the predictions is important in many safety-critical applications. However, modern deep neural networks are often overconfident for their incorrect predictions. Recently, many calibration methods have been proposed to alleviate the overconfidence problem. With calibrated confidence, a primary and practical purpose is to detect misclassification errors by filtering out low-confidence predictions (known as failure prediction). In this paper, we find a general, widely-existed but actually-neglected phenomenon that most confidence calibration methods are useless "},"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":"2303.02970","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2023-03-06T08:54:18Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"5422f32b62da58f4c49d8b0b7f9dfde6ca93911d41862d0801aa49c29bcccb08","abstract_canon_sha256":"5544452953bae3308b643a9dfd4f06727b16dfaa417d1e86710890f76d48c7d2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:48:17.715572Z","signature_b64":"Em7mA0o+CKXNg5DEg8/Q53LnGjA4uO2FOOvBzmeAvMiBCUQ4ES+mMd1sRqBkAe+F69DcJy01UR6tt4kQOdoeBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"442f9c409fb095a9fb84b067d478f83151c835afc36d5e19aa1618b9748e52ad","last_reissued_at":"2026-07-05T05:48:17.715168Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:48:17.715168Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Rethinking Confidence Calibration for Failure Prediction","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Cheng-Lin Liu, Fei Zhu, Xu-Yao Zhang, Zhen Cheng","submitted_at":"2023-03-06T08:54:18Z","abstract_excerpt":"Reliable confidence estimation for the predictions is important in many safety-critical applications. However, modern deep neural networks are often overconfident for their incorrect predictions. Recently, many calibration methods have been proposed to alleviate the overconfidence problem. With calibrated confidence, a primary and practical purpose is to detect misclassification errors by filtering out low-confidence predictions (known as failure prediction). In this paper, we find a general, widely-existed but actually-neglected phenomenon that most confidence calibration methods are useless "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2303.02970","kind":"arxiv","version":1},"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/2303.02970/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":"2303.02970","created_at":"2026-07-05T05:48:17.715224+00:00"},{"alias_kind":"arxiv_version","alias_value":"2303.02970v1","created_at":"2026-07-05T05:48:17.715224+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2303.02970","created_at":"2026-07-05T05:48:17.715224+00:00"},{"alias_kind":"pith_short_12","alias_value":"IQXZYQE7WCK2","created_at":"2026-07-05T05:48:17.715224+00:00"},{"alias_kind":"pith_short_16","alias_value":"IQXZYQE7WCK2T64E","created_at":"2026-07-05T05:48:17.715224+00:00"},{"alias_kind":"pith_short_8","alias_value":"IQXZYQE7","created_at":"2026-07-05T05:48:17.715224+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2502.05275","citing_title":"Interpretable Failure Detection with Human-Level Concepts","ref_index":44,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IQXZYQE7WCK2T64EWBT5I6HYGF","json":"https://pith.science/pith/IQXZYQE7WCK2T64EWBT5I6HYGF.json","graph_json":"https://pith.science/api/pith-number/IQXZYQE7WCK2T64EWBT5I6HYGF/graph.json","events_json":"https://pith.science/api/pith-number/IQXZYQE7WCK2T64EWBT5I6HYGF/events.json","paper":"https://pith.science/paper/IQXZYQE7"},"agent_actions":{"view_html":"https://pith.science/pith/IQXZYQE7WCK2T64EWBT5I6HYGF","download_json":"https://pith.science/pith/IQXZYQE7WCK2T64EWBT5I6HYGF.json","view_paper":"https://pith.science/paper/IQXZYQE7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2303.02970&json=true","fetch_graph":"https://pith.science/api/pith-number/IQXZYQE7WCK2T64EWBT5I6HYGF/graph.json","fetch_events":"https://pith.science/api/pith-number/IQXZYQE7WCK2T64EWBT5I6HYGF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IQXZYQE7WCK2T64EWBT5I6HYGF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IQXZYQE7WCK2T64EWBT5I6HYGF/action/storage_attestation","attest_author":"https://pith.science/pith/IQXZYQE7WCK2T64EWBT5I6HYGF/action/author_attestation","sign_citation":"https://pith.science/pith/IQXZYQE7WCK2T64EWBT5I6HYGF/action/citation_signature","submit_replication":"https://pith.science/pith/IQXZYQE7WCK2T64EWBT5I6HYGF/action/replication_record"}},"created_at":"2026-07-05T05:48:17.715224+00:00","updated_at":"2026-07-05T05:48:17.715224+00:00"}