{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MN3AQONROT6WCQ4TCWKUNI5IFF","short_pith_number":"pith:MN3AQONR","schema_version":"1.0","canonical_sha256":"63760839b174fd614393159546a3a82946645e933fad1ac6a22318c45a3c4159","source":{"kind":"arxiv","id":"2506.16407","version":1},"attestation_state":"computed","paper":{"title":"Robustness Evaluation of OCR-based Visual Document Understanding under Multi-Modal Adversarial Attacks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Dong Nguyen Tien, Dung D. Le","submitted_at":"2025-06-19T15:38:31Z","abstract_excerpt":"Visual Document Understanding (VDU) systems have achieved strong performance in information extraction by integrating textual, layout, and visual signals. However, their robustness under realistic adversarial perturbations remains insufficiently explored. We introduce the first unified framework for generating and evaluating multi-modal adversarial attacks on OCR-based VDU models. Our method covers six gradient-based layout attack scenarios, incorporating manipulations of OCR bounding boxes, pixels, and texts across both word and line granularities, with constraints on layout perturbation budg"},"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":"2506.16407","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-06-19T15:38:31Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"692e8c91932012ad08661d018a6ae56d8a193327c04ce25f10e8276668d974cb","abstract_canon_sha256":"fb156b7b26f59fdba26c80b7443c06ff9d68ba906bc665a5e7625eb12b443426"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:24:36.039713Z","signature_b64":"A94BrMA+JAB0bmH11IuChPQcjusFdhrNFWnU03byvtpPXL6pUfDE9nwPZXrsy/5xE6zTJptALaimmdbdBjFqAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"63760839b174fd614393159546a3a82946645e933fad1ac6a22318c45a3c4159","last_reissued_at":"2026-07-05T11:24:36.039195Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:24:36.039195Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Robustness Evaluation of OCR-based Visual Document Understanding under Multi-Modal Adversarial Attacks","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Dong Nguyen Tien, Dung D. Le","submitted_at":"2025-06-19T15:38:31Z","abstract_excerpt":"Visual Document Understanding (VDU) systems have achieved strong performance in information extraction by integrating textual, layout, and visual signals. However, their robustness under realistic adversarial perturbations remains insufficiently explored. We introduce the first unified framework for generating and evaluating multi-modal adversarial attacks on OCR-based VDU models. Our method covers six gradient-based layout attack scenarios, incorporating manipulations of OCR bounding boxes, pixels, and texts across both word and line granularities, with constraints on layout perturbation budg"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.16407","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/2506.16407/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":"2506.16407","created_at":"2026-07-05T11:24:36.039250+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.16407v1","created_at":"2026-07-05T11:24:36.039250+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.16407","created_at":"2026-07-05T11:24:36.039250+00:00"},{"alias_kind":"pith_short_12","alias_value":"MN3AQONROT6W","created_at":"2026-07-05T11:24:36.039250+00:00"},{"alias_kind":"pith_short_16","alias_value":"MN3AQONROT6WCQ4T","created_at":"2026-07-05T11:24:36.039250+00:00"},{"alias_kind":"pith_short_8","alias_value":"MN3AQONR","created_at":"2026-07-05T11:24:36.039250+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/MN3AQONROT6WCQ4TCWKUNI5IFF","json":"https://pith.science/pith/MN3AQONROT6WCQ4TCWKUNI5IFF.json","graph_json":"https://pith.science/api/pith-number/MN3AQONROT6WCQ4TCWKUNI5IFF/graph.json","events_json":"https://pith.science/api/pith-number/MN3AQONROT6WCQ4TCWKUNI5IFF/events.json","paper":"https://pith.science/paper/MN3AQONR"},"agent_actions":{"view_html":"https://pith.science/pith/MN3AQONROT6WCQ4TCWKUNI5IFF","download_json":"https://pith.science/pith/MN3AQONROT6WCQ4TCWKUNI5IFF.json","view_paper":"https://pith.science/paper/MN3AQONR","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.16407&json=true","fetch_graph":"https://pith.science/api/pith-number/MN3AQONROT6WCQ4TCWKUNI5IFF/graph.json","fetch_events":"https://pith.science/api/pith-number/MN3AQONROT6WCQ4TCWKUNI5IFF/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MN3AQONROT6WCQ4TCWKUNI5IFF/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MN3AQONROT6WCQ4TCWKUNI5IFF/action/storage_attestation","attest_author":"https://pith.science/pith/MN3AQONROT6WCQ4TCWKUNI5IFF/action/author_attestation","sign_citation":"https://pith.science/pith/MN3AQONROT6WCQ4TCWKUNI5IFF/action/citation_signature","submit_replication":"https://pith.science/pith/MN3AQONROT6WCQ4TCWKUNI5IFF/action/replication_record"}},"created_at":"2026-07-05T11:24:36.039250+00:00","updated_at":"2026-07-05T11:24:36.039250+00:00"}