{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2023:75AMPV2WF3MECSVYXL3V3UYSDQ","short_pith_number":"pith:75AMPV2W","canonical_record":{"source":{"id":"2301.12429","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-01-29T11:53:55Z","cross_cats_sorted":[],"title_canon_sha256":"e79458a762be34d7824ddd1599a1e11d20021db9f3e9ddd77dbda7abe2774051","abstract_canon_sha256":"569080bf3da89047ed07bc8852c51346c6ee0d8199558a3fbc5ef7e15d0fc10b"},"schema_version":"1.0"},"canonical_sha256":"ff40c7d7562ed8414ab8baf75dd3121c3a0281565d8237f9b262918f57206ce1","source":{"kind":"arxiv","id":"2301.12429","version":3},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2301.12429","created_at":"2026-07-05T11:52:50Z"},{"alias_kind":"arxiv_version","alias_value":"2301.12429v3","created_at":"2026-07-05T11:52:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2301.12429","created_at":"2026-07-05T11:52:50Z"},{"alias_kind":"pith_short_12","alias_value":"75AMPV2WF3ME","created_at":"2026-07-05T11:52:50Z"},{"alias_kind":"pith_short_16","alias_value":"75AMPV2WF3MECSVY","created_at":"2026-07-05T11:52:50Z"},{"alias_kind":"pith_short_8","alias_value":"75AMPV2W","created_at":"2026-07-05T11:52:50Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2023:75AMPV2WF3MECSVYXL3V3UYSDQ","target":"record","payload":{"canonical_record":{"source":{"id":"2301.12429","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-01-29T11:53:55Z","cross_cats_sorted":[],"title_canon_sha256":"e79458a762be34d7824ddd1599a1e11d20021db9f3e9ddd77dbda7abe2774051","abstract_canon_sha256":"569080bf3da89047ed07bc8852c51346c6ee0d8199558a3fbc5ef7e15d0fc10b"},"schema_version":"1.0"},"canonical_sha256":"ff40c7d7562ed8414ab8baf75dd3121c3a0281565d8237f9b262918f57206ce1","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:52:50.984042Z","signature_b64":"b3Dgigql84b2sG/YhPYaJ9WSrydth/UMEiDOfYolllyWdxNc3Nc5+jKsQzMHnO/fkT2shmizNxh6CCObAP3ABw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ff40c7d7562ed8414ab8baf75dd3121c3a0281565d8237f9b262918f57206ce1","last_reissued_at":"2026-07-05T11:52:50.983599Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:52:50.983599Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2301.12429","source_version":3,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T11:52:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"aa9XFVHvTELdkCqaLi3yQ0cyhLNUbPWWDumAgFiUDgCBiC5xQe7T2yP8TnGdROJs2Ttc/XZRB6TvL3dUnuzIAQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T15:21:55.944079Z"},"content_sha256":"b48441c30f3701d86d1dcbf6758c014c0334c9d933c173ac2c45312d36fecb25","schema_version":"1.0","event_id":"sha256:b48441c30f3701d86d1dcbf6758c014c0334c9d933c173ac2c45312d36fecb25"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2023:75AMPV2WF3MECSVYXL3V3UYSDQ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Debiased Fine-Tuning for Vision-language Models by Prompt Regularization","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Beier Zhu, Hanwang Zhang, Minhoe Hur, Saeil Lee, Yulei Niu","submitted_at":"2023-01-29T11:53:55Z","abstract_excerpt":"We present a new paradigm for fine-tuning large-scale visionlanguage pre-trained models on downstream task, dubbed Prompt Regularization (ProReg). Different from traditional fine-tuning which easily overfits to the downstream task data, ProReg uses the prediction by prompting the pretrained model to regularize the fine-tuning. The motivation is: by prompting the large model \"a photo of a [CLASS]\", the fil-lin answer is only dependent on the pretraining encyclopedic knowledge while independent of the task data distribution, which is usually biased. Specifically, given a training sample predicti"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2301.12429","kind":"arxiv","version":3},"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/2301.12429/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":"2026-07-05T11:52:50Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"CZkbMGQIEXyTzhYE+26pBJzjnnRfJtoW50DM3AkklpR05IF4H4yWpw6UThPLnp+7OefgfntwgFZ0bjLpTa5uAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-09T15:21:55.945188Z"},"content_sha256":"9caf2b1043d6c92540d0d300dabfa00da39958f2f2632da05a4d4b51ba063f76","schema_version":"1.0","event_id":"sha256:9caf2b1043d6c92540d0d300dabfa00da39958f2f2632da05a4d4b51ba063f76"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/75AMPV2WF3MECSVYXL3V3UYSDQ/bundle.json","state_url":"https://pith.science/pith/75AMPV2WF3MECSVYXL3V3UYSDQ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/75AMPV2WF3MECSVYXL3V3UYSDQ/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":"2026-08-09T15:21:55Z","links":{"resolver":"https://pith.science/pith/75AMPV2WF3MECSVYXL3V3UYSDQ","bundle":"https://pith.science/pith/75AMPV2WF3MECSVYXL3V3UYSDQ/bundle.json","state":"https://pith.science/pith/75AMPV2WF3MECSVYXL3V3UYSDQ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/75AMPV2WF3MECSVYXL3V3UYSDQ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2023:75AMPV2WF3MECSVYXL3V3UYSDQ","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":"569080bf3da89047ed07bc8852c51346c6ee0d8199558a3fbc5ef7e15d0fc10b","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-01-29T11:53:55Z","title_canon_sha256":"e79458a762be34d7824ddd1599a1e11d20021db9f3e9ddd77dbda7abe2774051"},"schema_version":"1.0","source":{"id":"2301.12429","kind":"arxiv","version":3}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2301.12429","created_at":"2026-07-05T11:52:50Z"},{"alias_kind":"arxiv_version","alias_value":"2301.12429v3","created_at":"2026-07-05T11:52:50Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2301.12429","created_at":"2026-07-05T11:52:50Z"},{"alias_kind":"pith_short_12","alias_value":"75AMPV2WF3ME","created_at":"2026-07-05T11:52:50Z"},{"alias_kind":"pith_short_16","alias_value":"75AMPV2WF3MECSVY","created_at":"2026-07-05T11:52:50Z"},{"alias_kind":"pith_short_8","alias_value":"75AMPV2W","created_at":"2026-07-05T11:52:50Z"}],"graph_snapshots":[{"event_id":"sha256:9caf2b1043d6c92540d0d300dabfa00da39958f2f2632da05a4d4b51ba063f76","target":"graph","created_at":"2026-07-05T11:52:50Z","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/2301.12429/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"We present a new paradigm for fine-tuning large-scale visionlanguage pre-trained models on downstream task, dubbed Prompt Regularization (ProReg). Different from traditional fine-tuning which easily overfits to the downstream task data, ProReg uses the prediction by prompting the pretrained model to regularize the fine-tuning. The motivation is: by prompting the large model \"a photo of a [CLASS]\", the fil-lin answer is only dependent on the pretraining encyclopedic knowledge while independent of the task data distribution, which is usually biased. Specifically, given a training sample predicti","authors_text":"Beier Zhu, Hanwang Zhang, Minhoe Hur, Saeil Lee, Yulei Niu","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-01-29T11:53:55Z","title":"Debiased Fine-Tuning for Vision-language Models by Prompt Regularization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2301.12429","kind":"arxiv","version":3},"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:b48441c30f3701d86d1dcbf6758c014c0334c9d933c173ac2c45312d36fecb25","target":"record","created_at":"2026-07-05T11:52:50Z","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":"569080bf3da89047ed07bc8852c51346c6ee0d8199558a3fbc5ef7e15d0fc10b","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-01-29T11:53:55Z","title_canon_sha256":"e79458a762be34d7824ddd1599a1e11d20021db9f3e9ddd77dbda7abe2774051"},"schema_version":"1.0","source":{"id":"2301.12429","kind":"arxiv","version":3}},"canonical_sha256":"ff40c7d7562ed8414ab8baf75dd3121c3a0281565d8237f9b262918f57206ce1","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"ff40c7d7562ed8414ab8baf75dd3121c3a0281565d8237f9b262918f57206ce1","first_computed_at":"2026-07-05T11:52:50.983599Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T11:52:50.983599Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"b3Dgigql84b2sG/YhPYaJ9WSrydth/UMEiDOfYolllyWdxNc3Nc5+jKsQzMHnO/fkT2shmizNxh6CCObAP3ABw==","signature_status":"signed_v1","signed_at":"2026-07-05T11:52:50.984042Z","signed_message":"canonical_sha256_bytes"},"source_id":"2301.12429","source_kind":"arxiv","source_version":3}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:b48441c30f3701d86d1dcbf6758c014c0334c9d933c173ac2c45312d36fecb25","sha256:9caf2b1043d6c92540d0d300dabfa00da39958f2f2632da05a4d4b51ba063f76"],"state_sha256":"59b842960db6e76d4ec31c2626b6ef368b47b0fb69ec567dcd46811ebc3e05d7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"reqJWWWvWcXg+j5ysatFRgoHCRNzfQKH+IHFin3xWt6gEhpQK74izX96K2XPMVQTA9j59UgGfxDo22L6X3doAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-09T15:21:55.952612Z","bundle_sha256":"45eecb724ae35430a9200acf559ae924edc8b7a5fe5f31703ec6245d9d43886b"}}