{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:2NDS3QTKMQMJMBZOH7EW366F3G","short_pith_number":"pith:2NDS3QTK","schema_version":"1.0","canonical_sha256":"d3472dc26a641896072e3fc96dfbc5d9be0d878e20dff93be0f1027c6cdb2395","source":{"kind":"arxiv","id":"2605.28023","version":1},"attestation_state":"computed","paper":{"title":"VCap: Hypergeometric Rewards for Weak-to-Strong Visual Captioning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.MM"],"primary_cat":"cs.CV","authors_text":"Bin Wen, Changyi Liu, Chun Yuan, Fan Yang, Han Li, Haonan Fan, Jinpeng Wang, Kaiyu Jiang, Tianke Zhang, Tingting Gao, Xingyu Lu, Xuanyu Zheng, Yancheng Long, Yankai Yang, Yi-Fan Zhang, Yiyang Fan","submitted_at":"2026-05-27T06:27:04Z","abstract_excerpt":"Visual captioning requires models to capture visual content faithfully while minimizing both omission and hallucination. As the dominant paradigm for captioning, MLLMs have achieved strong performance through scaling and high-quality data. Recently, RL has emerged as a key route to driving MLLMs toward higher precision and broader coverage, however, existing reward designs for captioning fail to provide fine-grained and reliable signals for factual verification, limiting their effectiveness. To address this, we propose VCap, a Witness-Adjudicator reward that pairs the reference caption (a witn"},"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":"2605.28023","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-05-27T06:27:04Z","cross_cats_sorted":["cs.AI","cs.CL","cs.MM"],"title_canon_sha256":"43e93e1ec97d573d56785b75df91547705a5408d85317444472072574595ca31","abstract_canon_sha256":"662a9ed723b1be9332368e4e58b7be0ceb1ab1b120bc19cf2ebd0a01204c3082"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-28T01:04:56.319826Z","signature_b64":"O79O6awZk9VlTLLxx63LLViC9QWzDzFP8gYsOxjXlKbKw8w8oKOHameLpFYb0AN03TxdZgrOfP5lhCMJJOfEDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d3472dc26a641896072e3fc96dfbc5d9be0d878e20dff93be0f1027c6cdb2395","last_reissued_at":"2026-05-28T01:04:56.319411Z","signature_status":"signed_v1","first_computed_at":"2026-05-28T01:04:56.319411Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"VCap: Hypergeometric Rewards for Weak-to-Strong Visual Captioning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CL","cs.MM"],"primary_cat":"cs.CV","authors_text":"Bin Wen, Changyi Liu, Chun Yuan, Fan Yang, Han Li, Haonan Fan, Jinpeng Wang, Kaiyu Jiang, Tianke Zhang, Tingting Gao, Xingyu Lu, Xuanyu Zheng, Yancheng Long, Yankai Yang, Yi-Fan Zhang, Yiyang Fan","submitted_at":"2026-05-27T06:27:04Z","abstract_excerpt":"Visual captioning requires models to capture visual content faithfully while minimizing both omission and hallucination. As the dominant paradigm for captioning, MLLMs have achieved strong performance through scaling and high-quality data. Recently, RL has emerged as a key route to driving MLLMs toward higher precision and broader coverage, however, existing reward designs for captioning fail to provide fine-grained and reliable signals for factual verification, limiting their effectiveness. To address this, we propose VCap, a Witness-Adjudicator reward that pairs the reference caption (a witn"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.28023","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/2605.28023/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":"2605.28023","created_at":"2026-05-28T01:04:56.319478+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.28023v1","created_at":"2026-05-28T01:04:56.319478+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.28023","created_at":"2026-05-28T01:04:56.319478+00:00"},{"alias_kind":"pith_short_12","alias_value":"2NDS3QTKMQMJ","created_at":"2026-05-28T01:04:56.319478+00:00"},{"alias_kind":"pith_short_16","alias_value":"2NDS3QTKMQMJMBZO","created_at":"2026-05-28T01:04:56.319478+00:00"},{"alias_kind":"pith_short_8","alias_value":"2NDS3QTK","created_at":"2026-05-28T01:04:56.319478+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/2NDS3QTKMQMJMBZOH7EW366F3G","json":"https://pith.science/pith/2NDS3QTKMQMJMBZOH7EW366F3G.json","graph_json":"https://pith.science/api/pith-number/2NDS3QTKMQMJMBZOH7EW366F3G/graph.json","events_json":"https://pith.science/api/pith-number/2NDS3QTKMQMJMBZOH7EW366F3G/events.json","paper":"https://pith.science/paper/2NDS3QTK"},"agent_actions":{"view_html":"https://pith.science/pith/2NDS3QTKMQMJMBZOH7EW366F3G","download_json":"https://pith.science/pith/2NDS3QTKMQMJMBZOH7EW366F3G.json","view_paper":"https://pith.science/paper/2NDS3QTK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.28023&json=true","fetch_graph":"https://pith.science/api/pith-number/2NDS3QTKMQMJMBZOH7EW366F3G/graph.json","fetch_events":"https://pith.science/api/pith-number/2NDS3QTKMQMJMBZOH7EW366F3G/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2NDS3QTKMQMJMBZOH7EW366F3G/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2NDS3QTKMQMJMBZOH7EW366F3G/action/storage_attestation","attest_author":"https://pith.science/pith/2NDS3QTKMQMJMBZOH7EW366F3G/action/author_attestation","sign_citation":"https://pith.science/pith/2NDS3QTKMQMJMBZOH7EW366F3G/action/citation_signature","submit_replication":"https://pith.science/pith/2NDS3QTKMQMJMBZOH7EW366F3G/action/replication_record"}},"created_at":"2026-05-28T01:04:56.319478+00:00","updated_at":"2026-05-28T01:04:56.319478+00:00"}