{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:TGIBLGEGCCLSRWERO7EYKTUOET","short_pith_number":"pith:TGIBLGEG","canonical_record":{"source":{"id":"2410.02681","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T17:06:21Z","cross_cats_sorted":[],"title_canon_sha256":"9a2b3c005e5ae29751e4d96a08ef15b9bfb3fe9d1ac0a362a00d675479d8c2a4","abstract_canon_sha256":"9162ce290545eaaa76aeca0b65af94350b441f5c856d27455419b0df2ecf1e39"},"schema_version":"1.0"},"canonical_sha256":"9990159886109728d89177c9854e8e24c8a88882706cbc2a5abdd813e2a58eb0","source":{"kind":"arxiv","id":"2410.02681","version":4},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.02681","created_at":"2026-07-05T11:11:35Z"},{"alias_kind":"arxiv_version","alias_value":"2410.02681v4","created_at":"2026-07-05T11:11:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.02681","created_at":"2026-07-05T11:11:35Z"},{"alias_kind":"pith_short_12","alias_value":"TGIBLGEGCCLS","created_at":"2026-07-05T11:11:35Z"},{"alias_kind":"pith_short_16","alias_value":"TGIBLGEGCCLSRWER","created_at":"2026-07-05T11:11:35Z"},{"alias_kind":"pith_short_8","alias_value":"TGIBLGEG","created_at":"2026-07-05T11:11:35Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:TGIBLGEGCCLSRWERO7EYKTUOET","target":"record","payload":{"canonical_record":{"source":{"id":"2410.02681","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T17:06:21Z","cross_cats_sorted":[],"title_canon_sha256":"9a2b3c005e5ae29751e4d96a08ef15b9bfb3fe9d1ac0a362a00d675479d8c2a4","abstract_canon_sha256":"9162ce290545eaaa76aeca0b65af94350b441f5c856d27455419b0df2ecf1e39"},"schema_version":"1.0"},"canonical_sha256":"9990159886109728d89177c9854e8e24c8a88882706cbc2a5abdd813e2a58eb0","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:11:35.227467Z","signature_b64":"Yj9TJp6UaG1r6TFb/4Kq7Wop5DcyTchCpN1n6r5JvzqY25vawnK6xmYEO2pTfonIC3e6i4/tAV6iRxfnyBCIBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9990159886109728d89177c9854e8e24c8a88882706cbc2a5abdd813e2a58eb0","last_reissued_at":"2026-07-05T11:11:35.226957Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:11:35.226957Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.02681","source_version":4,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:11:35Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"iWfzpA3QC6JAFtnf/bik8aAHlnee76rO7EbGbqxelw5U7hLzLnxtgzyLTbxK+mBPEW0vntmx7GYgVvz2xfQQAA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T07:42:49.276109Z"},"content_sha256":"761237a05cd367e2ee26ee4b5004b33a6d1153a3ce0881d56e6ed38dbd49caa1","schema_version":"1.0","event_id":"sha256:761237a05cd367e2ee26ee4b5004b33a6d1153a3ce0881d56e6ed38dbd49caa1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:TGIBLGEGCCLSRWERO7EYKTUOET","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Understanding and Mitigating Miscalibration in Prompt Tuning for Vision-Language Models","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Hongxin Wei, Shuoyuan Wang, Yixuan Li","submitted_at":"2024-10-03T17:06:21Z","abstract_excerpt":"Confidence calibration is critical for the safe deployment of machine learning models in the real world. However, such issue in vision-language models like CLIP, particularly after fine-tuning, has not been fully addressed. In this work, we demonstrate that existing prompt tuning methods usually lead to a trade-off of calibration between base and new classes: the cross-entropy loss in CoOp causes overconfidence in new classes by increasing textual label divergence, whereas the regularization of KgCoOp maintains the confidence level but results in underconfidence in base classes due to the impr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.02681","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/2410.02681/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:11:35Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"oqh/v9bUv3xl4N3NmfKfocDhrPJiwUDLTwF0IwJQw/lVMoXJ/OqTc5LzpGjptc/Moxi0qgb42KORhBrQlTqPBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T07:42:49.276610Z"},"content_sha256":"fbb1d59d88265ce803928b6d54bf15f7c598b807ecfd855d1da1b578b1340e4e","schema_version":"1.0","event_id":"sha256:fbb1d59d88265ce803928b6d54bf15f7c598b807ecfd855d1da1b578b1340e4e"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/TGIBLGEGCCLSRWERO7EYKTUOET/bundle.json","state_url":"https://pith.science/pith/TGIBLGEGCCLSRWERO7EYKTUOET/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/TGIBLGEGCCLSRWERO7EYKTUOET/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-08T07:42:49Z","links":{"resolver":"https://pith.science/pith/TGIBLGEGCCLSRWERO7EYKTUOET","bundle":"https://pith.science/pith/TGIBLGEGCCLSRWERO7EYKTUOET/bundle.json","state":"https://pith.science/pith/TGIBLGEGCCLSRWERO7EYKTUOET/state.json","well_known_bundle":"https://pith.science/.well-known/pith/TGIBLGEGCCLSRWERO7EYKTUOET/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:TGIBLGEGCCLSRWERO7EYKTUOET","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"9162ce290545eaaa76aeca0b65af94350b441f5c856d27455419b0df2ecf1e39","cross_cats_sorted":[],"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T17:06:21Z","title_canon_sha256":"9a2b3c005e5ae29751e4d96a08ef15b9bfb3fe9d1ac0a362a00d675479d8c2a4"},"schema_version":"1.0","source":{"id":"2410.02681","kind":"arxiv","version":4}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.02681","created_at":"2026-07-05T11:11:35Z"},{"alias_kind":"arxiv_version","alias_value":"2410.02681v4","created_at":"2026-07-05T11:11:35Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.02681","created_at":"2026-07-05T11:11:35Z"},{"alias_kind":"pith_short_12","alias_value":"TGIBLGEGCCLS","created_at":"2026-07-05T11:11:35Z"},{"alias_kind":"pith_short_16","alias_value":"TGIBLGEGCCLSRWER","created_at":"2026-07-05T11:11:35Z"},{"alias_kind":"pith_short_8","alias_value":"TGIBLGEG","created_at":"2026-07-05T11:11:35Z"}],"graph_snapshots":[{"event_id":"sha256:fbb1d59d88265ce803928b6d54bf15f7c598b807ecfd855d1da1b578b1340e4e","target":"graph","created_at":"2026-07-05T11:11:35Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2410.02681/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Confidence calibration is critical for the safe deployment of machine learning models in the real world. However, such issue in vision-language models like CLIP, particularly after fine-tuning, has not been fully addressed. In this work, we demonstrate that existing prompt tuning methods usually lead to a trade-off of calibration between base and new classes: the cross-entropy loss in CoOp causes overconfidence in new classes by increasing textual label divergence, whereas the regularization of KgCoOp maintains the confidence level but results in underconfidence in base classes due to the impr","authors_text":"Hongxin Wei, Shuoyuan Wang, Yixuan Li","cross_cats":[],"headline":"","license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T17:06:21Z","title":"Understanding and Mitigating Miscalibration in Prompt Tuning for Vision-Language Models"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.02681","kind":"arxiv","version":4},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:761237a05cd367e2ee26ee4b5004b33a6d1153a3ce0881d56e6ed38dbd49caa1","target":"record","created_at":"2026-07-05T11:11:35Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"9162ce290545eaaa76aeca0b65af94350b441f5c856d27455419b0df2ecf1e39","cross_cats_sorted":[],"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.LG","submitted_at":"2024-10-03T17:06:21Z","title_canon_sha256":"9a2b3c005e5ae29751e4d96a08ef15b9bfb3fe9d1ac0a362a00d675479d8c2a4"},"schema_version":"1.0","source":{"id":"2410.02681","kind":"arxiv","version":4}},"canonical_sha256":"9990159886109728d89177c9854e8e24c8a88882706cbc2a5abdd813e2a58eb0","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"9990159886109728d89177c9854e8e24c8a88882706cbc2a5abdd813e2a58eb0","first_computed_at":"2026-07-05T11:11:35.226957Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:11:35.226957Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"Yj9TJp6UaG1r6TFb/4Kq7Wop5DcyTchCpN1n6r5JvzqY25vawnK6xmYEO2pTfonIC3e6i4/tAV6iRxfnyBCIBw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:11:35.227467Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.02681","source_kind":"arxiv","source_version":4}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:761237a05cd367e2ee26ee4b5004b33a6d1153a3ce0881d56e6ed38dbd49caa1","sha256:fbb1d59d88265ce803928b6d54bf15f7c598b807ecfd855d1da1b578b1340e4e"],"state_sha256":"1ba7470ef3b8143b45b89863d4a127213d73f8b1a360f1c912ecd9be8a8d64a5"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+Fq4jXV8LYlIV2WNYmpBHs8LSzig32Aege7Swt4GNsJgO8A7mhoaIRSz9dLeaUHt7W3XOIEFIuLCi7emjwviAA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T07:42:49.281603Z","bundle_sha256":"68c5a71b89bd30383dc4ceff668163de095a6c949ed66af58af65f7645af8e5c"}}