{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:FRSWWE7GZTHZNB4DZHTOIY5WCS","short_pith_number":"pith:FRSWWE7G","canonical_record":{"source":{"id":"2401.02460","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-01-04T08:39:13Z","cross_cats_sorted":[],"title_canon_sha256":"6a77b13add1698520ca1709ecd3765af59e8e5a7250f3060bade44cb112568e1","abstract_canon_sha256":"ac9a56833ddcfb75aebc82342fee19a9531cd019fae09dbe9e074645e82b33c5"},"schema_version":"1.0"},"canonical_sha256":"2c656b13e6cccf968783c9e6e463b614994fb8442b2c3788ed75a73921a3d56c","source":{"kind":"arxiv","id":"2401.02460","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.02460","created_at":"2026-07-05T08:04:17Z"},{"alias_kind":"arxiv_version","alias_value":"2401.02460v2","created_at":"2026-07-05T08:04:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.02460","created_at":"2026-07-05T08:04:17Z"},{"alias_kind":"pith_short_12","alias_value":"FRSWWE7GZTHZ","created_at":"2026-07-05T08:04:17Z"},{"alias_kind":"pith_short_16","alias_value":"FRSWWE7GZTHZNB4D","created_at":"2026-07-05T08:04:17Z"},{"alias_kind":"pith_short_8","alias_value":"FRSWWE7G","created_at":"2026-07-05T08:04:17Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:FRSWWE7GZTHZNB4DZHTOIY5WCS","target":"record","payload":{"canonical_record":{"source":{"id":"2401.02460","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-01-04T08:39:13Z","cross_cats_sorted":[],"title_canon_sha256":"6a77b13add1698520ca1709ecd3765af59e8e5a7250f3060bade44cb112568e1","abstract_canon_sha256":"ac9a56833ddcfb75aebc82342fee19a9531cd019fae09dbe9e074645e82b33c5"},"schema_version":"1.0"},"canonical_sha256":"2c656b13e6cccf968783c9e6e463b614994fb8442b2c3788ed75a73921a3d56c","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:04:17.305831Z","signature_b64":"JqgKo0UAbJD7ngDtegVp9FI4pb7sD/PgVrzA58KAAJ41QXuEQxoidoOwpnvRdchm4tGHnU5olqKUhVvDXZNDAA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2c656b13e6cccf968783c9e6e463b614994fb8442b2c3788ed75a73921a3d56c","last_reissued_at":"2026-07-05T08:04:17.305255Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:04:17.305255Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2401.02460","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T08:04:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"vQX6M3hFjEbG+6+A+yhNMycKbKZzXSde1FvXE+7e1j14o6IoOwhZncNv6JscCcpbS0RagTeLYxJmPWLnmIfLDw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T12:15:11.737989Z"},"content_sha256":"4eef4f5794ddd244ecb673b5ddd113796784a0a45a2fc4041fd8f9ffb5e8db17","schema_version":"1.0","event_id":"sha256:4eef4f5794ddd244ecb673b5ddd113796784a0a45a2fc4041fd8f9ffb5e8db17"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:FRSWWE7GZTHZNB4DZHTOIY5WCS","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Improved Zero-Shot Classification by Adapting VLMs with Text Descriptions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Grant Van Horn, Oindrila Saha, Subhransu Maji","submitted_at":"2024-01-04T08:39:13Z","abstract_excerpt":"The zero-shot performance of existing vision-language models (VLMs) such as CLIP is limited by the availability of large-scale, aligned image and text datasets in specific domains. In this work, we leverage two complementary sources of information -- descriptions of categories generated by large language models (LLMs) and abundant, fine-grained image classification datasets -- to improve the zero-shot classification performance of VLMs across fine-grained domains. On the technical side, we develop methods to train VLMs with this \"bag-level\" image-text supervision. We find that simply using the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.02460","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/2401.02460/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":"2026-07-05T08:04:17Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"uv32QqQFiMrJVdsUWzXYIV/MzmYhcqX+ofQFKLzMHIHz7LWYC/eZ29SHvLEYQb5U6VP9jHxRsG63aVIXNK1+Ag==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-08T12:15:11.738488Z"},"content_sha256":"645358808dcffe2b62d46c2f0215772980abc0f541f69604c8779ac0997145bb","schema_version":"1.0","event_id":"sha256:645358808dcffe2b62d46c2f0215772980abc0f541f69604c8779ac0997145bb"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/FRSWWE7GZTHZNB4DZHTOIY5WCS/bundle.json","state_url":"https://pith.science/pith/FRSWWE7GZTHZNB4DZHTOIY5WCS/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/FRSWWE7GZTHZNB4DZHTOIY5WCS/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":"2026-08-08T12:15:11Z","links":{"resolver":"https://pith.science/pith/FRSWWE7GZTHZNB4DZHTOIY5WCS","bundle":"https://pith.science/pith/FRSWWE7GZTHZNB4DZHTOIY5WCS/bundle.json","state":"https://pith.science/pith/FRSWWE7GZTHZNB4DZHTOIY5WCS/state.json","well_known_bundle":"https://pith.science/.well-known/pith/FRSWWE7GZTHZNB4DZHTOIY5WCS/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:FRSWWE7GZTHZNB4DZHTOIY5WCS","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":"ac9a56833ddcfb75aebc82342fee19a9531cd019fae09dbe9e074645e82b33c5","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-01-04T08:39:13Z","title_canon_sha256":"6a77b13add1698520ca1709ecd3765af59e8e5a7250f3060bade44cb112568e1"},"schema_version":"1.0","source":{"id":"2401.02460","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2401.02460","created_at":"2026-07-05T08:04:17Z"},{"alias_kind":"arxiv_version","alias_value":"2401.02460v2","created_at":"2026-07-05T08:04:17Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2401.02460","created_at":"2026-07-05T08:04:17Z"},{"alias_kind":"pith_short_12","alias_value":"FRSWWE7GZTHZ","created_at":"2026-07-05T08:04:17Z"},{"alias_kind":"pith_short_16","alias_value":"FRSWWE7GZTHZNB4D","created_at":"2026-07-05T08:04:17Z"},{"alias_kind":"pith_short_8","alias_value":"FRSWWE7G","created_at":"2026-07-05T08:04:17Z"}],"graph_snapshots":[{"event_id":"sha256:645358808dcffe2b62d46c2f0215772980abc0f541f69604c8779ac0997145bb","target":"graph","created_at":"2026-07-05T08:04:17Z","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/2401.02460/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The zero-shot performance of existing vision-language models (VLMs) such as CLIP is limited by the availability of large-scale, aligned image and text datasets in specific domains. In this work, we leverage two complementary sources of information -- descriptions of categories generated by large language models (LLMs) and abundant, fine-grained image classification datasets -- to improve the zero-shot classification performance of VLMs across fine-grained domains. On the technical side, we develop methods to train VLMs with this \"bag-level\" image-text supervision. We find that simply using the","authors_text":"Grant Van Horn, Oindrila Saha, Subhransu Maji","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-01-04T08:39:13Z","title":"Improved Zero-Shot Classification by Adapting VLMs with Text Descriptions"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2401.02460","kind":"arxiv","version":2},"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":"sha256:4eef4f5794ddd244ecb673b5ddd113796784a0a45a2fc4041fd8f9ffb5e8db17","target":"record","created_at":"2026-07-05T08:04:17Z","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":"ac9a56833ddcfb75aebc82342fee19a9531cd019fae09dbe9e074645e82b33c5","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-01-04T08:39:13Z","title_canon_sha256":"6a77b13add1698520ca1709ecd3765af59e8e5a7250f3060bade44cb112568e1"},"schema_version":"1.0","source":{"id":"2401.02460","kind":"arxiv","version":2}},"canonical_sha256":"2c656b13e6cccf968783c9e6e463b614994fb8442b2c3788ed75a73921a3d56c","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"2c656b13e6cccf968783c9e6e463b614994fb8442b2c3788ed75a73921a3d56c","first_computed_at":"2026-07-05T08:04:17.305255Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:04:17.305255Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"JqgKo0UAbJD7ngDtegVp9FI4pb7sD/PgVrzA58KAAJ41QXuEQxoidoOwpnvRdchm4tGHnU5olqKUhVvDXZNDAA==","signature_status":"signed_v1","signed_at":"2026-07-05T08:04:17.305831Z","signed_message":"canonical_sha256_bytes"},"source_id":"2401.02460","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:4eef4f5794ddd244ecb673b5ddd113796784a0a45a2fc4041fd8f9ffb5e8db17","sha256:645358808dcffe2b62d46c2f0215772980abc0f541f69604c8779ac0997145bb"],"state_sha256":"6740c9509514d28a3213fc17c5fa18c5853907a7d69f77359aa2708880fec129"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"tpaVdx3O3FgqhqC/jPdwshK3uKglNnA63OB2DuRiwPaWzwrRX6gfq2XJgomN2BDzxhNxN4YeYea6Y9sIfEzQDQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-08T12:15:11.742700Z","bundle_sha256":"501f848b6b8aa7c9910d355a2871d352d286ce2dc8f468ee7a880c6ba25071df"}}