{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:HPXWSEPXQRVZKWQZI2VWFNCUD2","short_pith_number":"pith:HPXWSEPX","schema_version":"1.0","canonical_sha256":"3bef6911f7846b955a1946ab62b4541e899241076fdd8eaad17ff865ea003203","source":{"kind":"arxiv","id":"2205.01883","version":1},"attestation_state":"computed","paper":{"title":"All You May Need for VQA are Image Captions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Doron Kukliansky, Idan Szpektor, Nan Ding, Radu Soricut, Soravit Changpinyo, Xi Chen","submitted_at":"2022-05-04T04:09:23Z","abstract_excerpt":"Visual Question Answering (VQA) has benefited from increasingly sophisticated models, but has not enjoyed the same level of engagement in terms of data creation. In this paper, we propose a method that automatically derives VQA examples at volume, by leveraging the abundance of existing image-caption annotations combined with neural models for textual question generation. We show that the resulting data is of high-quality. VQA models trained on our data improve state-of-the-art zero-shot accuracy by double digits and achieve a level of robustness that lacks in the same model trained on human-a"},"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":"2205.01883","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2022-05-04T04:09:23Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"e3b1676179e2ef0f91347882e3d4cca9393c9fd58dae9db194d4c766d62b0965","abstract_canon_sha256":"39bea77ffecb0224fd3baac6ddbe5fbd21447141f0aabd94dc21146e30c8f374"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:20:23.231261Z","signature_b64":"GH87PS/FkHfBFR6tb1dbEO8Z5zRgkD2iL6IOco/mx9WqEa5xlrhlEuGTA7JR4fDajDOcmqXtn6Hf95MmhYmDCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3bef6911f7846b955a1946ab62b4541e899241076fdd8eaad17ff865ea003203","last_reissued_at":"2026-07-05T04:20:23.230711Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:20:23.230711Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"All You May Need for VQA are Image Captions","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.CV","authors_text":"Doron Kukliansky, Idan Szpektor, Nan Ding, Radu Soricut, Soravit Changpinyo, Xi Chen","submitted_at":"2022-05-04T04:09:23Z","abstract_excerpt":"Visual Question Answering (VQA) has benefited from increasingly sophisticated models, but has not enjoyed the same level of engagement in terms of data creation. In this paper, we propose a method that automatically derives VQA examples at volume, by leveraging the abundance of existing image-caption annotations combined with neural models for textual question generation. We show that the resulting data is of high-quality. VQA models trained on our data improve state-of-the-art zero-shot accuracy by double digits and achieve a level of robustness that lacks in the same model trained on human-a"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2205.01883","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/2205.01883/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":"2205.01883","created_at":"2026-07-05T04:20:23.230779+00:00"},{"alias_kind":"arxiv_version","alias_value":"2205.01883v1","created_at":"2026-07-05T04:20:23.230779+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2205.01883","created_at":"2026-07-05T04:20:23.230779+00:00"},{"alias_kind":"pith_short_12","alias_value":"HPXWSEPXQRVZ","created_at":"2026-07-05T04:20:23.230779+00:00"},{"alias_kind":"pith_short_16","alias_value":"HPXWSEPXQRVZKWQZ","created_at":"2026-07-05T04:20:23.230779+00:00"},{"alias_kind":"pith_short_8","alias_value":"HPXWSEPX","created_at":"2026-07-05T04:20:23.230779+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":2,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19584","citing_title":"Language-Instructed Vision Embeddings for Controllable and Generalizable Perception","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2303.03378","citing_title":"PaLM-E: An Embodied Multimodal Language Model","ref_index":6,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/HPXWSEPXQRVZKWQZI2VWFNCUD2","json":"https://pith.science/pith/HPXWSEPXQRVZKWQZI2VWFNCUD2.json","graph_json":"https://pith.science/api/pith-number/HPXWSEPXQRVZKWQZI2VWFNCUD2/graph.json","events_json":"https://pith.science/api/pith-number/HPXWSEPXQRVZKWQZI2VWFNCUD2/events.json","paper":"https://pith.science/paper/HPXWSEPX"},"agent_actions":{"view_html":"https://pith.science/pith/HPXWSEPXQRVZKWQZI2VWFNCUD2","download_json":"https://pith.science/pith/HPXWSEPXQRVZKWQZI2VWFNCUD2.json","view_paper":"https://pith.science/paper/HPXWSEPX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2205.01883&json=true","fetch_graph":"https://pith.science/api/pith-number/HPXWSEPXQRVZKWQZI2VWFNCUD2/graph.json","fetch_events":"https://pith.science/api/pith-number/HPXWSEPXQRVZKWQZI2VWFNCUD2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/HPXWSEPXQRVZKWQZI2VWFNCUD2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/HPXWSEPXQRVZKWQZI2VWFNCUD2/action/storage_attestation","attest_author":"https://pith.science/pith/HPXWSEPXQRVZKWQZI2VWFNCUD2/action/author_attestation","sign_citation":"https://pith.science/pith/HPXWSEPXQRVZKWQZI2VWFNCUD2/action/citation_signature","submit_replication":"https://pith.science/pith/HPXWSEPXQRVZKWQZI2VWFNCUD2/action/replication_record"}},"created_at":"2026-07-05T04:20:23.230779+00:00","updated_at":"2026-07-05T04:20:23.230779+00:00"}