{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:7GCOPCZ7Q5PB74XQFFBDFAGIUB","short_pith_number":"pith:7GCOPCZ7","schema_version":"1.0","canonical_sha256":"f984e78b3f875e1ff2f029423280c8a05746d3524a09ac99c083174604eba65f","source":{"kind":"arxiv","id":"2508.08589","version":1},"attestation_state":"computed","paper":{"title":"DocThinker: Explainable Multimodal Large Language Models with Rule-based Reinforcement Learning for Document Understanding","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Wenwen Yu, Xiang Bai, Yuliang Liu, Zhibo Yang","submitted_at":"2025-08-12T03:06:55Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) have demonstrated remarkable capabilities in document understanding. However, their reasoning processes remain largely black-box, making it difficult to ensure reliability and trustworthiness, especially in high-stakes domains such as legal, financial, and medical document analysis. Existing methods use fixed Chain-of-Thought (CoT) reasoning with supervised fine-tuning (SFT) but suffer from catastrophic forgetting, poor adaptability, and limited generalization across domain tasks. In this paper, we propose DocThinker, a rule-based Reinforcement Learning"},"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":"2508.08589","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"cs.CV","submitted_at":"2025-08-12T03:06:55Z","cross_cats_sorted":[],"title_canon_sha256":"ff5c805f3c44620371008bcea4b040df7fe5d4a6fa81e2619bbd3e90e775dc27","abstract_canon_sha256":"8cd260881596599a53c9b9b9d1bf0a48a0869ce035a426f454f5afd6b3d5c4be"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:52:30.921348Z","signature_b64":"6pFMhiLJBsM6k/1CaNAjRU4fxjK0xUNKVkaDVNc00oHC9c0iw4n+zNT3i5TTZ98aLmHaIpDJpqoHNj2GuUThAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f984e78b3f875e1ff2f029423280c8a05746d3524a09ac99c083174604eba65f","last_reissued_at":"2026-07-05T11:52:30.920854Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:52:30.920854Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"DocThinker: Explainable Multimodal Large Language Models with Rule-based Reinforcement Learning for Document Understanding","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Wenwen Yu, Xiang Bai, Yuliang Liu, Zhibo Yang","submitted_at":"2025-08-12T03:06:55Z","abstract_excerpt":"Multimodal Large Language Models (MLLMs) have demonstrated remarkable capabilities in document understanding. However, their reasoning processes remain largely black-box, making it difficult to ensure reliability and trustworthiness, especially in high-stakes domains such as legal, financial, and medical document analysis. Existing methods use fixed Chain-of-Thought (CoT) reasoning with supervised fine-tuning (SFT) but suffer from catastrophic forgetting, poor adaptability, and limited generalization across domain tasks. In this paper, we propose DocThinker, a rule-based Reinforcement Learning"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2508.08589","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/2508.08589/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":"2508.08589","created_at":"2026-07-05T11:52:30.920914+00:00"},{"alias_kind":"arxiv_version","alias_value":"2508.08589v1","created_at":"2026-07-05T11:52:30.920914+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2508.08589","created_at":"2026-07-05T11:52:30.920914+00:00"},{"alias_kind":"pith_short_12","alias_value":"7GCOPCZ7Q5PB","created_at":"2026-07-05T11:52:30.920914+00:00"},{"alias_kind":"pith_short_16","alias_value":"7GCOPCZ7Q5PB74XQ","created_at":"2026-07-05T11:52:30.920914+00:00"},{"alias_kind":"pith_short_8","alias_value":"7GCOPCZ7","created_at":"2026-07-05T11:52:30.920914+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/7GCOPCZ7Q5PB74XQFFBDFAGIUB","json":"https://pith.science/pith/7GCOPCZ7Q5PB74XQFFBDFAGIUB.json","graph_json":"https://pith.science/api/pith-number/7GCOPCZ7Q5PB74XQFFBDFAGIUB/graph.json","events_json":"https://pith.science/api/pith-number/7GCOPCZ7Q5PB74XQFFBDFAGIUB/events.json","paper":"https://pith.science/paper/7GCOPCZ7"},"agent_actions":{"view_html":"https://pith.science/pith/7GCOPCZ7Q5PB74XQFFBDFAGIUB","download_json":"https://pith.science/pith/7GCOPCZ7Q5PB74XQFFBDFAGIUB.json","view_paper":"https://pith.science/paper/7GCOPCZ7","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2508.08589&json=true","fetch_graph":"https://pith.science/api/pith-number/7GCOPCZ7Q5PB74XQFFBDFAGIUB/graph.json","fetch_events":"https://pith.science/api/pith-number/7GCOPCZ7Q5PB74XQFFBDFAGIUB/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/7GCOPCZ7Q5PB74XQFFBDFAGIUB/action/timestamp_anchor","attest_storage":"https://pith.science/pith/7GCOPCZ7Q5PB74XQFFBDFAGIUB/action/storage_attestation","attest_author":"https://pith.science/pith/7GCOPCZ7Q5PB74XQFFBDFAGIUB/action/author_attestation","sign_citation":"https://pith.science/pith/7GCOPCZ7Q5PB74XQFFBDFAGIUB/action/citation_signature","submit_replication":"https://pith.science/pith/7GCOPCZ7Q5PB74XQFFBDFAGIUB/action/replication_record"}},"created_at":"2026-07-05T11:52:30.920914+00:00","updated_at":"2026-07-05T11:52:30.920914+00:00"}