{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:YQPJVUXCX3YG75YZ7EFR3LXIWT","short_pith_number":"pith:YQPJVUXC","schema_version":"1.0","canonical_sha256":"c41e9ad2e2bef06ff719f90b1daee8b4c3d6d4495827a256668e959e9ee43239","source":{"kind":"arxiv","id":"2412.03561","version":1},"attestation_state":"computed","paper":{"title":"FLAIR: VLM with Fine-grained Language-informed Image Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Mariana-Iuliana Georgescu, Rui Xiao, Sanghwan Kim, Stephan Alaniz, Zeynep Akata","submitted_at":"2024-12-04T18:56:04Z","abstract_excerpt":"CLIP has shown impressive results in aligning images and texts at scale. However, its ability to capture detailed visual features remains limited because CLIP matches images and texts at a global level. To address this issue, we propose FLAIR, Fine-grained Language-informed Image Representations, an approach that utilizes long and detailed image descriptions to learn localized image embeddings. By sampling diverse sub-captions that describe fine-grained details about an image, we train our vision-language model to produce not only global embeddings but also text-specific image representations."},"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":"2412.03561","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-12-04T18:56:04Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"ccad5bce95b500c83ccccba8f1940aa0afb0e151eac56442886e979f6ffac97b","abstract_canon_sha256":"95841116a673536a8d6773a4de64230a47e9f616c90c265764ee2e6f027d61da"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:44:31.318357Z","signature_b64":"v9M3nhAbF3X6m778puJ+gXfmt+eHO+UhmgXAKKIdeO/EofkfHisefeA6tnoqaT096lxC6vcQG86drVLyZ/p2Dw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c41e9ad2e2bef06ff719f90b1daee8b4c3d6d4495827a256668e959e9ee43239","last_reissued_at":"2026-07-05T09:44:31.317778Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:44:31.317778Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FLAIR: VLM with Fine-grained Language-informed Image Representations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Mariana-Iuliana Georgescu, Rui Xiao, Sanghwan Kim, Stephan Alaniz, Zeynep Akata","submitted_at":"2024-12-04T18:56:04Z","abstract_excerpt":"CLIP has shown impressive results in aligning images and texts at scale. However, its ability to capture detailed visual features remains limited because CLIP matches images and texts at a global level. To address this issue, we propose FLAIR, Fine-grained Language-informed Image Representations, an approach that utilizes long and detailed image descriptions to learn localized image embeddings. By sampling diverse sub-captions that describe fine-grained details about an image, we train our vision-language model to produce not only global embeddings but also text-specific image representations."},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2412.03561","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/2412.03561/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":"2412.03561","created_at":"2026-07-05T09:44:31.317847+00:00"},{"alias_kind":"arxiv_version","alias_value":"2412.03561v1","created_at":"2026-07-05T09:44:31.317847+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2412.03561","created_at":"2026-07-05T09:44:31.317847+00:00"},{"alias_kind":"pith_short_12","alias_value":"YQPJVUXCX3YG","created_at":"2026-07-05T09:44:31.317847+00:00"},{"alias_kind":"pith_short_16","alias_value":"YQPJVUXCX3YG75YZ","created_at":"2026-07-05T09:44:31.317847+00:00"},{"alias_kind":"pith_short_8","alias_value":"YQPJVUXC","created_at":"2026-07-05T09:44:31.317847+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.11659","citing_title":"Reviving In-domain Fine-tuning Methods for Source-Free Cross-domain Few-shot Learning","ref_index":37,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YQPJVUXCX3YG75YZ7EFR3LXIWT","json":"https://pith.science/pith/YQPJVUXCX3YG75YZ7EFR3LXIWT.json","graph_json":"https://pith.science/api/pith-number/YQPJVUXCX3YG75YZ7EFR3LXIWT/graph.json","events_json":"https://pith.science/api/pith-number/YQPJVUXCX3YG75YZ7EFR3LXIWT/events.json","paper":"https://pith.science/paper/YQPJVUXC"},"agent_actions":{"view_html":"https://pith.science/pith/YQPJVUXCX3YG75YZ7EFR3LXIWT","download_json":"https://pith.science/pith/YQPJVUXCX3YG75YZ7EFR3LXIWT.json","view_paper":"https://pith.science/paper/YQPJVUXC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2412.03561&json=true","fetch_graph":"https://pith.science/api/pith-number/YQPJVUXCX3YG75YZ7EFR3LXIWT/graph.json","fetch_events":"https://pith.science/api/pith-number/YQPJVUXCX3YG75YZ7EFR3LXIWT/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YQPJVUXCX3YG75YZ7EFR3LXIWT/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YQPJVUXCX3YG75YZ7EFR3LXIWT/action/storage_attestation","attest_author":"https://pith.science/pith/YQPJVUXCX3YG75YZ7EFR3LXIWT/action/author_attestation","sign_citation":"https://pith.science/pith/YQPJVUXCX3YG75YZ7EFR3LXIWT/action/citation_signature","submit_replication":"https://pith.science/pith/YQPJVUXCX3YG75YZ7EFR3LXIWT/action/replication_record"}},"created_at":"2026-07-05T09:44:31.317847+00:00","updated_at":"2026-07-05T09:44:31.317847+00:00"}