{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:66P3ZLUZGD4WOMRBSUCWPJSORN","short_pith_number":"pith:66P3ZLUZ","schema_version":"1.0","canonical_sha256":"f79fbcae9930f9673221950567a64e8b4dadb3167f769b175b55efe3f44bf4e3","source":{"kind":"arxiv","id":"2307.09233","version":3},"attestation_state":"computed","paper":{"title":"Distilling Knowledge from Text-to-Image Generative Models Improves Visio-Linguistic Reasoning in CLIP","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Daniela Massiceti, Maziar Sanjabi, Samyadeep Basu, Shell Xu Hu, Soheil Feizi","submitted_at":"2023-07-18T13:10:11Z","abstract_excerpt":"Image-text contrastive models like CLIP have wide applications in zero-shot classification, image-text retrieval, and transfer learning. However, they often struggle on compositional visio-linguistic tasks (e.g., attribute-binding or object-relationships) where their performance is no better than random chance. To address this, we introduce SDS-CLIP, a lightweight and sample-efficient distillation method to enhance CLIP's compositional visio-linguistic reasoning. Our approach fine-tunes CLIP using a distillation objective borrowed from large text-to-image generative models like Stable-Diffusio"},"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":"2307.09233","kind":"arxiv","version":3},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2023-07-18T13:10:11Z","cross_cats_sorted":[],"title_canon_sha256":"9145cb4857bff0ab2776de182f0c05ac23e0fa40909b2a918956514d3a8556e8","abstract_canon_sha256":"b16cdab3160143b52c7e46349cb1cab81a1f8d4ed8026e9fa4e3bcfad0595146"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:38:33.563149Z","signature_b64":"Qonkk9kvaGSvYR5tnRkCJE6gLjjkOxcMMCgpqbeNlSRKER+d+bbtMUSDVfRscR0Yuus9vzARcKvVmEE/rpf+Bg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f79fbcae9930f9673221950567a64e8b4dadb3167f769b175b55efe3f44bf4e3","last_reissued_at":"2026-07-05T08:38:33.562653Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:38:33.562653Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Distilling Knowledge from Text-to-Image Generative Models Improves Visio-Linguistic Reasoning in CLIP","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Daniela Massiceti, Maziar Sanjabi, Samyadeep Basu, Shell Xu Hu, Soheil Feizi","submitted_at":"2023-07-18T13:10:11Z","abstract_excerpt":"Image-text contrastive models like CLIP have wide applications in zero-shot classification, image-text retrieval, and transfer learning. However, they often struggle on compositional visio-linguistic tasks (e.g., attribute-binding or object-relationships) where their performance is no better than random chance. To address this, we introduce SDS-CLIP, a lightweight and sample-efficient distillation method to enhance CLIP's compositional visio-linguistic reasoning. Our approach fine-tunes CLIP using a distillation objective borrowed from large text-to-image generative models like Stable-Diffusio"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2307.09233","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/2307.09233/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":"2307.09233","created_at":"2026-07-05T08:38:33.562714+00:00"},{"alias_kind":"arxiv_version","alias_value":"2307.09233v3","created_at":"2026-07-05T08:38:33.562714+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2307.09233","created_at":"2026-07-05T08:38:33.562714+00:00"},{"alias_kind":"pith_short_12","alias_value":"66P3ZLUZGD4W","created_at":"2026-07-05T08:38:33.562714+00:00"},{"alias_kind":"pith_short_16","alias_value":"66P3ZLUZGD4WOMRB","created_at":"2026-07-05T08:38:33.562714+00:00"},{"alias_kind":"pith_short_8","alias_value":"66P3ZLUZ","created_at":"2026-07-05T08:38:33.562714+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.12335","citing_title":"All in One: A Unified Synthetic Data Pipeline for Multimodal Video Understanding","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/66P3ZLUZGD4WOMRBSUCWPJSORN","json":"https://pith.science/pith/66P3ZLUZGD4WOMRBSUCWPJSORN.json","graph_json":"https://pith.science/api/pith-number/66P3ZLUZGD4WOMRBSUCWPJSORN/graph.json","events_json":"https://pith.science/api/pith-number/66P3ZLUZGD4WOMRBSUCWPJSORN/events.json","paper":"https://pith.science/paper/66P3ZLUZ"},"agent_actions":{"view_html":"https://pith.science/pith/66P3ZLUZGD4WOMRBSUCWPJSORN","download_json":"https://pith.science/pith/66P3ZLUZGD4WOMRBSUCWPJSORN.json","view_paper":"https://pith.science/paper/66P3ZLUZ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2307.09233&json=true","fetch_graph":"https://pith.science/api/pith-number/66P3ZLUZGD4WOMRBSUCWPJSORN/graph.json","fetch_events":"https://pith.science/api/pith-number/66P3ZLUZGD4WOMRBSUCWPJSORN/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/66P3ZLUZGD4WOMRBSUCWPJSORN/action/timestamp_anchor","attest_storage":"https://pith.science/pith/66P3ZLUZGD4WOMRBSUCWPJSORN/action/storage_attestation","attest_author":"https://pith.science/pith/66P3ZLUZGD4WOMRBSUCWPJSORN/action/author_attestation","sign_citation":"https://pith.science/pith/66P3ZLUZGD4WOMRBSUCWPJSORN/action/citation_signature","submit_replication":"https://pith.science/pith/66P3ZLUZGD4WOMRBSUCWPJSORN/action/replication_record"}},"created_at":"2026-07-05T08:38:33.562714+00:00","updated_at":"2026-07-05T08:38:33.562714+00:00"}