{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:FZAZWZ4MAA4OSUBXXUETYUSFDB","short_pith_number":"pith:FZAZWZ4M","schema_version":"1.0","canonical_sha256":"2e419b678c0038e95037bd093c52451843ce6f111e65c6406913b6738c7a08c3","source":{"kind":"arxiv","id":"2305.02549","version":2},"attestation_state":"computed","paper":{"title":"FormNetV2: Multimodal Graph Contrastive Learning for Form Document Information Extraction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chen-Yu Lee, Chun-Liang Li, Guolong Su, Hao Zhang, Joshua Ainslie, Kihyuk Sohn, Nan Hua, Nikolai Glushnev, Renshen Wang, Shangbang Long, Siyang Qin, Timothy Dozat, Tomas Pfister, Vincent Perot, Xiang Zhang, Yasuhisa Fujii","submitted_at":"2023-05-04T05:02:04Z","abstract_excerpt":"The recent advent of self-supervised pre-training techniques has led to a surge in the use of multimodal learning in form document understanding. However, existing approaches that extend the mask language modeling to other modalities require careful multi-task tuning, complex reconstruction target designs, or additional pre-training data. In FormNetV2, we introduce a centralized multimodal graph contrastive learning strategy to unify self-supervised pre-training for all modalities in one loss. The graph contrastive objective maximizes the agreement of multimodal representations, providing a na"},"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":"2305.02549","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CL","submitted_at":"2023-05-04T05:02:04Z","cross_cats_sorted":["cs.CV","cs.LG"],"title_canon_sha256":"35c8b82eba4f2e94c0674c18ec922a953c07c82b41fbfb8d498e018f13d1df66","abstract_canon_sha256":"629f86f20985fcc7b6bf57d7d60d08780488624a9984df9c6b7f81d0e9a79167"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:19:58.671457Z","signature_b64":"iSZwNM/3pl6KqzSTxyqRYPDOp2yK6Wi3e9fk1ZMQBfMyl7KG4YrnaksYoFBlrM7L+jEfm4+WjymNWjVuS5u9Cw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2e419b678c0038e95037bd093c52451843ce6f111e65c6406913b6738c7a08c3","last_reissued_at":"2026-07-05T06:19:58.670917Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:19:58.670917Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"FormNetV2: Multimodal Graph Contrastive Learning for Form Document Information Extraction","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","cs.LG"],"primary_cat":"cs.CL","authors_text":"Chen-Yu Lee, Chun-Liang Li, Guolong Su, Hao Zhang, Joshua Ainslie, Kihyuk Sohn, Nan Hua, Nikolai Glushnev, Renshen Wang, Shangbang Long, Siyang Qin, Timothy Dozat, Tomas Pfister, Vincent Perot, Xiang Zhang, Yasuhisa Fujii","submitted_at":"2023-05-04T05:02:04Z","abstract_excerpt":"The recent advent of self-supervised pre-training techniques has led to a surge in the use of multimodal learning in form document understanding. However, existing approaches that extend the mask language modeling to other modalities require careful multi-task tuning, complex reconstruction target designs, or additional pre-training data. In FormNetV2, we introduce a centralized multimodal graph contrastive learning strategy to unify self-supervised pre-training for all modalities in one loss. The graph contrastive objective maximizes the agreement of multimodal representations, providing a na"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.02549","kind":"arxiv","version":2},"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/2305.02549/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":"2305.02549","created_at":"2026-07-05T06:19:58.670991+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.02549v2","created_at":"2026-07-05T06:19:58.670991+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.02549","created_at":"2026-07-05T06:19:58.670991+00:00"},{"alias_kind":"pith_short_12","alias_value":"FZAZWZ4MAA4O","created_at":"2026-07-05T06:19:58.670991+00:00"},{"alias_kind":"pith_short_16","alias_value":"FZAZWZ4MAA4OSUBX","created_at":"2026-07-05T06:19:58.670991+00:00"},{"alias_kind":"pith_short_8","alias_value":"FZAZWZ4M","created_at":"2026-07-05T06:19:58.670991+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/FZAZWZ4MAA4OSUBXXUETYUSFDB","json":"https://pith.science/pith/FZAZWZ4MAA4OSUBXXUETYUSFDB.json","graph_json":"https://pith.science/api/pith-number/FZAZWZ4MAA4OSUBXXUETYUSFDB/graph.json","events_json":"https://pith.science/api/pith-number/FZAZWZ4MAA4OSUBXXUETYUSFDB/events.json","paper":"https://pith.science/paper/FZAZWZ4M"},"agent_actions":{"view_html":"https://pith.science/pith/FZAZWZ4MAA4OSUBXXUETYUSFDB","download_json":"https://pith.science/pith/FZAZWZ4MAA4OSUBXXUETYUSFDB.json","view_paper":"https://pith.science/paper/FZAZWZ4M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.02549&json=true","fetch_graph":"https://pith.science/api/pith-number/FZAZWZ4MAA4OSUBXXUETYUSFDB/graph.json","fetch_events":"https://pith.science/api/pith-number/FZAZWZ4MAA4OSUBXXUETYUSFDB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/FZAZWZ4MAA4OSUBXXUETYUSFDB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/FZAZWZ4MAA4OSUBXXUETYUSFDB/action/storage_attestation","attest_author":"https://pith.science/pith/FZAZWZ4MAA4OSUBXXUETYUSFDB/action/author_attestation","sign_citation":"https://pith.science/pith/FZAZWZ4MAA4OSUBXXUETYUSFDB/action/citation_signature","submit_replication":"https://pith.science/pith/FZAZWZ4MAA4OSUBXXUETYUSFDB/action/replication_record"}},"created_at":"2026-07-05T06:19:58.670991+00:00","updated_at":"2026-07-05T06:19:58.670991+00:00"}