{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:5P7LY5ZVK566CEGT5LJIP5APPQ","short_pith_number":"pith:5P7LY5ZV","schema_version":"1.0","canonical_sha256":"ebfebc7735577de110d3ead287f40f7c2545cf8335900b7af03db112144dbfd4","source":{"kind":"arxiv","id":"2404.00419","version":1},"attestation_state":"computed","paper":{"title":"Do Vision-Language Models Understand Compound Nouns?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Dinesh Manocha, Sonal Kumar, Sreyan Ghosh, S Sakshi, Utkarsh Tyagi","submitted_at":"2024-03-30T16:54:45Z","abstract_excerpt":"Open-vocabulary vision-language models (VLMs) like CLIP, trained using contrastive loss, have emerged as a promising new paradigm for text-to-image retrieval. However, do VLMs understand compound nouns (CNs) (e.g., lab coat) as well as they understand nouns (e.g., lab)? We curate Compun, a novel benchmark with 400 unique and commonly used CNs, to evaluate the effectiveness of VLMs in interpreting CNs. The Compun benchmark challenges a VLM for text-to-image retrieval where, given a text prompt with a CN, the task is to select the correct image that shows the CN among a pair of distractor images"},"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":"2404.00419","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-03-30T16:54:45Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"177a71d9113904ef6b24c7342269d462057af4d08fd456b114a3598e75404635","abstract_canon_sha256":"bc3a420f6188d2618e3894717be2af779711e2a4be2125c3e615e7c3c2ad4bbc"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:02:33.022465Z","signature_b64":"hXTK+PB7aQNZXtC4llEKnv6QYvwx4mOpkYQmNmS0MBlvOIYiULn6DwytsxU5Ge4tBQvQuf8L90MD7V1dSwKcBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"ebfebc7735577de110d3ead287f40f7c2545cf8335900b7af03db112144dbfd4","last_reissued_at":"2026-07-05T08:02:33.021997Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:02:33.021997Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Do Vision-Language Models Understand Compound Nouns?","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Dinesh Manocha, Sonal Kumar, Sreyan Ghosh, S Sakshi, Utkarsh Tyagi","submitted_at":"2024-03-30T16:54:45Z","abstract_excerpt":"Open-vocabulary vision-language models (VLMs) like CLIP, trained using contrastive loss, have emerged as a promising new paradigm for text-to-image retrieval. However, do VLMs understand compound nouns (CNs) (e.g., lab coat) as well as they understand nouns (e.g., lab)? We curate Compun, a novel benchmark with 400 unique and commonly used CNs, to evaluate the effectiveness of VLMs in interpreting CNs. The Compun benchmark challenges a VLM for text-to-image retrieval where, given a text prompt with a CN, the task is to select the correct image that shows the CN among a pair of distractor images"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.00419","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/2404.00419/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":"2404.00419","created_at":"2026-07-05T08:02:33.022052+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.00419v1","created_at":"2026-07-05T08:02:33.022052+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.00419","created_at":"2026-07-05T08:02:33.022052+00:00"},{"alias_kind":"pith_short_12","alias_value":"5P7LY5ZVK566","created_at":"2026-07-05T08:02:33.022052+00:00"},{"alias_kind":"pith_short_16","alias_value":"5P7LY5ZVK566CEGT","created_at":"2026-07-05T08:02:33.022052+00:00"},{"alias_kind":"pith_short_8","alias_value":"5P7LY5ZV","created_at":"2026-07-05T08:02:33.022052+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2412.08158","citing_title":"How Vision-Language Tasks Benefit from Large Pre-trained Models: A Survey","ref_index":179,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/5P7LY5ZVK566CEGT5LJIP5APPQ","json":"https://pith.science/pith/5P7LY5ZVK566CEGT5LJIP5APPQ.json","graph_json":"https://pith.science/api/pith-number/5P7LY5ZVK566CEGT5LJIP5APPQ/graph.json","events_json":"https://pith.science/api/pith-number/5P7LY5ZVK566CEGT5LJIP5APPQ/events.json","paper":"https://pith.science/paper/5P7LY5ZV"},"agent_actions":{"view_html":"https://pith.science/pith/5P7LY5ZVK566CEGT5LJIP5APPQ","download_json":"https://pith.science/pith/5P7LY5ZVK566CEGT5LJIP5APPQ.json","view_paper":"https://pith.science/paper/5P7LY5ZV","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.00419&json=true","fetch_graph":"https://pith.science/api/pith-number/5P7LY5ZVK566CEGT5LJIP5APPQ/graph.json","fetch_events":"https://pith.science/api/pith-number/5P7LY5ZVK566CEGT5LJIP5APPQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/5P7LY5ZVK566CEGT5LJIP5APPQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/5P7LY5ZVK566CEGT5LJIP5APPQ/action/storage_attestation","attest_author":"https://pith.science/pith/5P7LY5ZVK566CEGT5LJIP5APPQ/action/author_attestation","sign_citation":"https://pith.science/pith/5P7LY5ZVK566CEGT5LJIP5APPQ/action/citation_signature","submit_replication":"https://pith.science/pith/5P7LY5ZVK566CEGT5LJIP5APPQ/action/replication_record"}},"created_at":"2026-07-05T08:02:33.022052+00:00","updated_at":"2026-07-05T08:02:33.022052+00:00"}