{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:IZ2XYB5XMSJTQSL5CKOX2L4JFW","short_pith_number":"pith:IZ2XYB5X","schema_version":"1.0","canonical_sha256":"46757c07b7649338497d129d7d2f892d9ccde60cb07f9f07cd8c56d9399456fe","source":{"kind":"arxiv","id":"2408.14339","version":1},"attestation_state":"computed","paper":{"title":"ConceptMix: A Compositional Image Generation Benchmark with Controllable Difficulty","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dingli Yu, Olga Russakovsky, Sanjeev Arora, Xindi Wu, Yangsibo Huang","submitted_at":"2024-08-26T15:08:12Z","abstract_excerpt":"Compositionality is a critical capability in Text-to-Image (T2I) models, as it reflects their ability to understand and combine multiple concepts from text descriptions. Existing evaluations of compositional capability rely heavily on human-designed text prompts or fixed templates, limiting their diversity and complexity, and yielding low discriminative power. We propose ConceptMix, a scalable, controllable, and customizable benchmark which automatically evaluates compositional generation ability of T2I models. This is done in two stages. First, ConceptMix generates the text prompts: concretel"},"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":"2408.14339","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-08-26T15:08:12Z","cross_cats_sorted":[],"title_canon_sha256":"dd143d495df0890ccc337ff0fe3a1328cfb720b91bd8f94ed043a6b148ba5d44","abstract_canon_sha256":"237fdb5e90e2d0c70435f1e853337c3e720704f9836e89fead11cb8aa9e8a176"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:59:19.465882Z","signature_b64":"rsOJBNfMkg1bn+ndV90QctFWHmioU+cFss02pocGjcI2to3wNrKqerK0DWXuCTtyhL+o9IOtMhYBG1rZP9viBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"46757c07b7649338497d129d7d2f892d9ccde60cb07f9f07cd8c56d9399456fe","last_reissued_at":"2026-07-05T08:59:19.465468Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:59:19.465468Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"ConceptMix: A Compositional Image Generation Benchmark with Controllable Difficulty","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Dingli Yu, Olga Russakovsky, Sanjeev Arora, Xindi Wu, Yangsibo Huang","submitted_at":"2024-08-26T15:08:12Z","abstract_excerpt":"Compositionality is a critical capability in Text-to-Image (T2I) models, as it reflects their ability to understand and combine multiple concepts from text descriptions. Existing evaluations of compositional capability rely heavily on human-designed text prompts or fixed templates, limiting their diversity and complexity, and yielding low discriminative power. We propose ConceptMix, a scalable, controllable, and customizable benchmark which automatically evaluates compositional generation ability of T2I models. This is done in two stages. First, ConceptMix generates the text prompts: concretel"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2408.14339","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/2408.14339/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":"2408.14339","created_at":"2026-07-05T08:59:19.465529+00:00"},{"alias_kind":"arxiv_version","alias_value":"2408.14339v1","created_at":"2026-07-05T08:59:19.465529+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2408.14339","created_at":"2026-07-05T08:59:19.465529+00:00"},{"alias_kind":"pith_short_12","alias_value":"IZ2XYB5XMSJT","created_at":"2026-07-05T08:59:19.465529+00:00"},{"alias_kind":"pith_short_16","alias_value":"IZ2XYB5XMSJTQSL5","created_at":"2026-07-05T08:59:19.465529+00:00"},{"alias_kind":"pith_short_8","alias_value":"IZ2XYB5X","created_at":"2026-07-05T08:59:19.465529+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2412.04300","citing_title":"T2I-FactualBench: Benchmarking the Factuality of Text-to-Image Models with Knowledge-Intensive Concepts","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2506.09082","citing_title":"AVA-Bench: Atomic Visual Ability Benchmark for Vision Foundation Models","ref_index":100,"is_internal_anchor":false},{"citing_arxiv_id":"2503.07265","citing_title":"WISE: A World Knowledge-Informed Semantic Evaluation for Text-to-Image Generation","ref_index":53,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/IZ2XYB5XMSJTQSL5CKOX2L4JFW","json":"https://pith.science/pith/IZ2XYB5XMSJTQSL5CKOX2L4JFW.json","graph_json":"https://pith.science/api/pith-number/IZ2XYB5XMSJTQSL5CKOX2L4JFW/graph.json","events_json":"https://pith.science/api/pith-number/IZ2XYB5XMSJTQSL5CKOX2L4JFW/events.json","paper":"https://pith.science/paper/IZ2XYB5X"},"agent_actions":{"view_html":"https://pith.science/pith/IZ2XYB5XMSJTQSL5CKOX2L4JFW","download_json":"https://pith.science/pith/IZ2XYB5XMSJTQSL5CKOX2L4JFW.json","view_paper":"https://pith.science/paper/IZ2XYB5X","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2408.14339&json=true","fetch_graph":"https://pith.science/api/pith-number/IZ2XYB5XMSJTQSL5CKOX2L4JFW/graph.json","fetch_events":"https://pith.science/api/pith-number/IZ2XYB5XMSJTQSL5CKOX2L4JFW/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/IZ2XYB5XMSJTQSL5CKOX2L4JFW/action/timestamp_anchor","attest_storage":"https://pith.science/pith/IZ2XYB5XMSJTQSL5CKOX2L4JFW/action/storage_attestation","attest_author":"https://pith.science/pith/IZ2XYB5XMSJTQSL5CKOX2L4JFW/action/author_attestation","sign_citation":"https://pith.science/pith/IZ2XYB5XMSJTQSL5CKOX2L4JFW/action/citation_signature","submit_replication":"https://pith.science/pith/IZ2XYB5XMSJTQSL5CKOX2L4JFW/action/replication_record"}},"created_at":"2026-07-05T08:59:19.465529+00:00","updated_at":"2026-07-05T08:59:19.465529+00:00"}