{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:XRQ3225LOLU7FHZWL3522Q3NUI","short_pith_number":"pith:XRQ3225L","schema_version":"1.0","canonical_sha256":"bc61bd6bab72e9f29f365efbad436da22b523f00739b4c740978cad6048082c3","source":{"kind":"arxiv","id":"2508.20783","version":1},"attestation_state":"computed","paper":{"title":"Evaluating Compositional Generalisation in VLMs and Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Beth Pearson, Bilal Boulbarss, Martha Lewis, Michael Wray","submitted_at":"2025-08-28T13:45:04Z","abstract_excerpt":"A fundamental aspect of the semantics of natural language is that novel meanings can be formed from the composition of previously known parts. Vision-language models (VLMs) have made significant progress in recent years, however, there is evidence that they are unable to perform this kind of composition. For example, given an image of a red cube and a blue cylinder, a VLM such as CLIP is likely to incorrectly label the image as a red cylinder or a blue cube, indicating it represents the image as a `bag-of-words' and fails to capture compositional semantics. Diffusion models have recently gaine"},"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":"2508.20783","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-08-28T13:45:04Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"8206cb50bf2c8664ca6c6cdc380bfec655441ae3b42390f1ee97ef1492c59eeb","abstract_canon_sha256":"9e7fa31391af33722632c75ec82fda712ac6ea2b7e7ccced042d5b2d4be7b9e5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T12:01:10.782837Z","signature_b64":"JGiUYlL84Q5DpXohAY0sdIb/osHilCKVj69qo+qpgoH8s365kSciSvmVr6VBMfVp5CD8m2lQ0uStuCxjU8+0Ag==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"bc61bd6bab72e9f29f365efbad436da22b523f00739b4c740978cad6048082c3","last_reissued_at":"2026-07-05T12:01:10.782293Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T12:01:10.782293Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Evaluating Compositional Generalisation in VLMs and Diffusion Models","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Beth Pearson, Bilal Boulbarss, Martha Lewis, Michael Wray","submitted_at":"2025-08-28T13:45:04Z","abstract_excerpt":"A fundamental aspect of the semantics of natural language is that novel meanings can be formed from the composition of previously known parts. Vision-language models (VLMs) have made significant progress in recent years, however, there is evidence that they are unable to perform this kind of composition. For example, given an image of a red cube and a blue cylinder, a VLM such as CLIP is likely to incorrectly label the image as a red cylinder or a blue cube, indicating it represents the image as a `bag-of-words' and fails to capture compositional semantics. Diffusion models have recently gaine"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.20783","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/2508.20783/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":"2508.20783","created_at":"2026-07-05T12:01:10.782365+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.20783v1","created_at":"2026-07-05T12:01:10.782365+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.20783","created_at":"2026-07-05T12:01:10.782365+00:00"},{"alias_kind":"pith_short_12","alias_value":"XRQ3225LOLU7","created_at":"2026-07-05T12:01:10.782365+00:00"},{"alias_kind":"pith_short_16","alias_value":"XRQ3225LOLU7FHZW","created_at":"2026-07-05T12:01:10.782365+00:00"},{"alias_kind":"pith_short_8","alias_value":"XRQ3225L","created_at":"2026-07-05T12:01:10.782365+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/XRQ3225LOLU7FHZWL3522Q3NUI","json":"https://pith.science/pith/XRQ3225LOLU7FHZWL3522Q3NUI.json","graph_json":"https://pith.science/api/pith-number/XRQ3225LOLU7FHZWL3522Q3NUI/graph.json","events_json":"https://pith.science/api/pith-number/XRQ3225LOLU7FHZWL3522Q3NUI/events.json","paper":"https://pith.science/paper/XRQ3225L"},"agent_actions":{"view_html":"https://pith.science/pith/XRQ3225LOLU7FHZWL3522Q3NUI","download_json":"https://pith.science/pith/XRQ3225LOLU7FHZWL3522Q3NUI.json","view_paper":"https://pith.science/paper/XRQ3225L","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.20783&json=true","fetch_graph":"https://pith.science/api/pith-number/XRQ3225LOLU7FHZWL3522Q3NUI/graph.json","fetch_events":"https://pith.science/api/pith-number/XRQ3225LOLU7FHZWL3522Q3NUI/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/XRQ3225LOLU7FHZWL3522Q3NUI/action/timestamp_anchor","attest_storage":"https://pith.science/pith/XRQ3225LOLU7FHZWL3522Q3NUI/action/storage_attestation","attest_author":"https://pith.science/pith/XRQ3225LOLU7FHZWL3522Q3NUI/action/author_attestation","sign_citation":"https://pith.science/pith/XRQ3225LOLU7FHZWL3522Q3NUI/action/citation_signature","submit_replication":"https://pith.science/pith/XRQ3225LOLU7FHZWL3522Q3NUI/action/replication_record"}},"created_at":"2026-07-05T12:01:10.782365+00:00","updated_at":"2026-07-05T12:01:10.782365+00:00"}