{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:4YZ3EUAYLRKXKUSEZU3VXYFY7H","short_pith_number":"pith:4YZ3EUAY","schema_version":"1.0","canonical_sha256":"e633b250185c55755244cd375be0b8f9cbe0cdfeacc1922670951439e8ab3a54","source":{"kind":"arxiv","id":"2506.01725","version":1},"attestation_state":"computed","paper":{"title":"VideoCap-R1: Enhancing MLLMs for Video Captioning via Structured Thinking","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Desen Meng, Jun Zhang, Limin Wang, Lingshu Zhang, Meng Zhang, Rui Huang, Xinhao Li, Yifan Xu, Yi Liu, Zhenpeng Huang, Zhilin Dai","submitted_at":"2025-06-02T14:30:09Z","abstract_excerpt":"While recent advances in reinforcement learning have significantly enhanced reasoning capabilities in large language models (LLMs), these techniques remain underexplored in multi-modal LLMs for video captioning. This paper presents the first systematic investigation of GRPO-based RL post-training for video MLLMs, with the goal of enhancing video MLLMs' capability of describing actions in videos. Specifically, we develop the VideoCap-R1, which is prompted to first perform structured thinking that analyzes video subjects with their attributes and actions before generating complete captions, supp"},"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":"2506.01725","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-06-02T14:30:09Z","cross_cats_sorted":[],"title_canon_sha256":"d792cef7173f8295fa67b0d0d6cd2ee4e12fadf8b5af55da0d183383d8575128","abstract_canon_sha256":"402caf08301a715448ea2c9badf3a6b6d429474522bf2de008249f63a9003a50"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:14:18.450946Z","signature_b64":"pQBns4PlbxpoAzzVfX3pufB95TAI0pLjQjK7JAhlQBu+rSfZXzE0LXtrHvRFZVR3nEivsoB03/mCRoGZUbvnAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e633b250185c55755244cd375be0b8f9cbe0cdfeacc1922670951439e8ab3a54","last_reissued_at":"2026-07-05T11:14:18.450499Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:14:18.450499Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VideoCap-R1: Enhancing MLLMs for Video Captioning via Structured Thinking","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Desen Meng, Jun Zhang, Limin Wang, Lingshu Zhang, Meng Zhang, Rui Huang, Xinhao Li, Yifan Xu, Yi Liu, Zhenpeng Huang, Zhilin Dai","submitted_at":"2025-06-02T14:30:09Z","abstract_excerpt":"While recent advances in reinforcement learning have significantly enhanced reasoning capabilities in large language models (LLMs), these techniques remain underexplored in multi-modal LLMs for video captioning. This paper presents the first systematic investigation of GRPO-based RL post-training for video MLLMs, with the goal of enhancing video MLLMs' capability of describing actions in videos. Specifically, we develop the VideoCap-R1, which is prompted to first perform structured thinking that analyzes video subjects with their attributes and actions before generating complete captions, supp"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2506.01725","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/2506.01725/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":"2506.01725","created_at":"2026-07-05T11:14:18.450563+00:00"},{"alias_kind":"arxiv_version","alias_value":"2506.01725v1","created_at":"2026-07-05T11:14:18.450563+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2506.01725","created_at":"2026-07-05T11:14:18.450563+00:00"},{"alias_kind":"pith_short_12","alias_value":"4YZ3EUAYLRKX","created_at":"2026-07-05T11:14:18.450563+00:00"},{"alias_kind":"pith_short_16","alias_value":"4YZ3EUAYLRKXKUSE","created_at":"2026-07-05T11:14:18.450563+00:00"},{"alias_kind":"pith_short_8","alias_value":"4YZ3EUAY","created_at":"2026-07-05T11:14:18.450563+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.24636","citing_title":"CineCap: Structured Reasoning with Spatio-Temporal Anchors for Cinematographic Video Captioning","ref_index":25,"is_internal_anchor":false},{"citing_arxiv_id":"2606.26196","citing_title":"From Structure to Synergy: A Survey of Vision-Language Perception Paradigm Evolution in Multimodal Large Language Models","ref_index":194,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07433","citing_title":"Watch, Remember, Reason: Human-View Video Understanding with MLLMs","ref_index":94,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28023","citing_title":"VCap: Hypergeometric Rewards for Weak-to-Strong Visual Captioning","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2511.11113","citing_title":"VIDEOP2R: Video Understanding from Perception to Reasoning","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2604.21718","citing_title":"Building a Precise Video Language with Human-AI Oversight","ref_index":44,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/4YZ3EUAYLRKXKUSEZU3VXYFY7H","json":"https://pith.science/pith/4YZ3EUAYLRKXKUSEZU3VXYFY7H.json","graph_json":"https://pith.science/api/pith-number/4YZ3EUAYLRKXKUSEZU3VXYFY7H/graph.json","events_json":"https://pith.science/api/pith-number/4YZ3EUAYLRKXKUSEZU3VXYFY7H/events.json","paper":"https://pith.science/paper/4YZ3EUAY"},"agent_actions":{"view_html":"https://pith.science/pith/4YZ3EUAYLRKXKUSEZU3VXYFY7H","download_json":"https://pith.science/pith/4YZ3EUAYLRKXKUSEZU3VXYFY7H.json","view_paper":"https://pith.science/paper/4YZ3EUAY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2506.01725&json=true","fetch_graph":"https://pith.science/api/pith-number/4YZ3EUAYLRKXKUSEZU3VXYFY7H/graph.json","fetch_events":"https://pith.science/api/pith-number/4YZ3EUAYLRKXKUSEZU3VXYFY7H/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/4YZ3EUAYLRKXKUSEZU3VXYFY7H/action/timestamp_anchor","attest_storage":"https://pith.science/pith/4YZ3EUAYLRKXKUSEZU3VXYFY7H/action/storage_attestation","attest_author":"https://pith.science/pith/4YZ3EUAYLRKXKUSEZU3VXYFY7H/action/author_attestation","sign_citation":"https://pith.science/pith/4YZ3EUAYLRKXKUSEZU3VXYFY7H/action/citation_signature","submit_replication":"https://pith.science/pith/4YZ3EUAYLRKXKUSEZU3VXYFY7H/action/replication_record"}},"created_at":"2026-07-05T11:14:18.450563+00:00","updated_at":"2026-07-05T11:14:18.450563+00:00"}