{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:ZOI2D4OC2U5XGVYQZGD4EGKWVV","short_pith_number":"pith:ZOI2D4OC","schema_version":"1.0","canonical_sha256":"cb91a1f1c2d53b735710c987c21956ad74f2d981b03bb0a778697c25aad35194","source":{"kind":"arxiv","id":"2210.15037","version":1},"attestation_state":"computed","paper":{"title":"Generalization Differences between End-to-End and Neuro-Symbolic Vision-Language Reasoning Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CL","authors_text":"Jesse Thomason, Robin Jia, Wang Zhu","submitted_at":"2022-10-26T21:11:47Z","abstract_excerpt":"For vision-and-language reasoning tasks, both fully connectionist, end-to-end methods and hybrid, neuro-symbolic methods have achieved high in-distribution performance. In which out-of-distribution settings does each paradigm excel? We investigate this question on both single-image and multi-image visual question-answering through four types of generalization tests: a novel segment-combine test for multi-image queries, contrast set, compositional generalization, and cross-benchmark transfer. Vision-and-language end-to-end trained systems exhibit sizeable performance drops across all these test"},"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":"2210.15037","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2022-10-26T21:11:47Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"a3620ffe81bf66d640354639069204afda3257c1a220f96d0da445dd551e71f3","abstract_canon_sha256":"271980eec36b1b7d4fe46e69bbdff3d4078adc4886a0927c3a0f1045df061c5e"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T05:11:01.301543Z","signature_b64":"m4/zQU9ydLZ8fsf3gFqEI2PacIPJLfsmJTqM/0sjhJ/lQA8xnE+D5MRJL4Bf4RFYfptlNObyFMYYPVvKFZQoAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cb91a1f1c2d53b735710c987c21956ad74f2d981b03bb0a778697c25aad35194","last_reissued_at":"2026-07-05T05:11:01.301131Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T05:11:01.301131Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Generalization Differences between End-to-End and Neuro-Symbolic Vision-Language Reasoning Systems","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.CL","authors_text":"Jesse Thomason, Robin Jia, Wang Zhu","submitted_at":"2022-10-26T21:11:47Z","abstract_excerpt":"For vision-and-language reasoning tasks, both fully connectionist, end-to-end methods and hybrid, neuro-symbolic methods have achieved high in-distribution performance. In which out-of-distribution settings does each paradigm excel? We investigate this question on both single-image and multi-image visual question-answering through four types of generalization tests: a novel segment-combine test for multi-image queries, contrast set, compositional generalization, and cross-benchmark transfer. Vision-and-language end-to-end trained systems exhibit sizeable performance drops across all these test"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2210.15037","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/2210.15037/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":"2210.15037","created_at":"2026-07-05T05:11:01.301196+00:00"},{"alias_kind":"arxiv_version","alias_value":"2210.15037v1","created_at":"2026-07-05T05:11:01.301196+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2210.15037","created_at":"2026-07-05T05:11:01.301196+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZOI2D4OC2U5X","created_at":"2026-07-05T05:11:01.301196+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZOI2D4OC2U5XGVYQ","created_at":"2026-07-05T05:11:01.301196+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZOI2D4OC","created_at":"2026-07-05T05:11:01.301196+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/ZOI2D4OC2U5XGVYQZGD4EGKWVV","json":"https://pith.science/pith/ZOI2D4OC2U5XGVYQZGD4EGKWVV.json","graph_json":"https://pith.science/api/pith-number/ZOI2D4OC2U5XGVYQZGD4EGKWVV/graph.json","events_json":"https://pith.science/api/pith-number/ZOI2D4OC2U5XGVYQZGD4EGKWVV/events.json","paper":"https://pith.science/paper/ZOI2D4OC"},"agent_actions":{"view_html":"https://pith.science/pith/ZOI2D4OC2U5XGVYQZGD4EGKWVV","download_json":"https://pith.science/pith/ZOI2D4OC2U5XGVYQZGD4EGKWVV.json","view_paper":"https://pith.science/paper/ZOI2D4OC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2210.15037&json=true","fetch_graph":"https://pith.science/api/pith-number/ZOI2D4OC2U5XGVYQZGD4EGKWVV/graph.json","fetch_events":"https://pith.science/api/pith-number/ZOI2D4OC2U5XGVYQZGD4EGKWVV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZOI2D4OC2U5XGVYQZGD4EGKWVV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZOI2D4OC2U5XGVYQZGD4EGKWVV/action/storage_attestation","attest_author":"https://pith.science/pith/ZOI2D4OC2U5XGVYQZGD4EGKWVV/action/author_attestation","sign_citation":"https://pith.science/pith/ZOI2D4OC2U5XGVYQZGD4EGKWVV/action/citation_signature","submit_replication":"https://pith.science/pith/ZOI2D4OC2U5XGVYQZGD4EGKWVV/action/replication_record"}},"created_at":"2026-07-05T05:11:01.301196+00:00","updated_at":"2026-07-05T05:11:01.301196+00:00"}