{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:U2QYX74FIHIC5AIK2BNY2NTVLP","short_pith_number":"pith:U2QYX74F","schema_version":"1.0","canonical_sha256":"a6a18bff8541d02e810ad05b8d36755bddcefff45391b92307e9875cc083023b","source":{"kind":"arxiv","id":"2505.07263","version":2},"attestation_state":"computed","paper":{"title":"Skywork-VL Reward: An Effective Reward Model for Multimodal Understanding and Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ai Jian, Jiangbo Pei, Peiyu Wang, Tianyidan Xie, Weijie Qiu, Wei Shen, Xiaokun Wang, Xuchen Song, Yahui Zhou, Yang Liu, Yi Peng, Yunzhuo Hao","submitted_at":"2025-05-12T06:23:08Z","abstract_excerpt":"We propose Skywork-VL Reward, a multimodal reward model that provides reward signals for both multimodal understanding and reasoning tasks. Our technical approach comprises two key components: First, we construct a large-scale multimodal preference dataset that covers a wide range of tasks and scenarios, with responses collected from both standard vision-language models (VLMs) and advanced VLM reasoners. Second, we design a reward model architecture based on Qwen2.5-VL-7B-Instruct, integrating a reward head and applying multi-stage fine-tuning using pairwise ranking loss on pairwise preference"},"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":"2505.07263","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2025-05-12T06:23:08Z","cross_cats_sorted":[],"title_canon_sha256":"a00e8a48e7f7fa9eb240d9309b564790f3efec260bcfff8cbd480184a45c56f9","abstract_canon_sha256":"1c785d4d3f1b66f666a7ad3e3faab84b2c84421610cec7919813936351ef9b80"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:18:21.678350Z","signature_b64":"WAfJcpRiUx6XpBotzOH0UqVU80D6h/f/X8/S/sEti4jsElnRutqhpqgqzajXsPlNgF42jzo1Xj9H1BferXuxDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a6a18bff8541d02e810ad05b8d36755bddcefff45391b92307e9875cc083023b","last_reissued_at":"2026-07-05T11:18:21.677846Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:18:21.677846Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Skywork-VL Reward: An Effective Reward Model for Multimodal Understanding and Reasoning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ai Jian, Jiangbo Pei, Peiyu Wang, Tianyidan Xie, Weijie Qiu, Wei Shen, Xiaokun Wang, Xuchen Song, Yahui Zhou, Yang Liu, Yi Peng, Yunzhuo Hao","submitted_at":"2025-05-12T06:23:08Z","abstract_excerpt":"We propose Skywork-VL Reward, a multimodal reward model that provides reward signals for both multimodal understanding and reasoning tasks. Our technical approach comprises two key components: First, we construct a large-scale multimodal preference dataset that covers a wide range of tasks and scenarios, with responses collected from both standard vision-language models (VLMs) and advanced VLM reasoners. Second, we design a reward model architecture based on Qwen2.5-VL-7B-Instruct, integrating a reward head and applying multi-stage fine-tuning using pairwise ranking loss on pairwise preference"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.07263","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/2505.07263/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":"2505.07263","created_at":"2026-07-05T11:18:21.677906+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.07263v2","created_at":"2026-07-05T11:18:21.677906+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.07263","created_at":"2026-07-05T11:18:21.677906+00:00"},{"alias_kind":"pith_short_12","alias_value":"U2QYX74FIHIC","created_at":"2026-07-05T11:18:21.677906+00:00"},{"alias_kind":"pith_short_16","alias_value":"U2QYX74FIHIC5AIK","created_at":"2026-07-05T11:18:21.677906+00:00"},{"alias_kind":"pith_short_8","alias_value":"U2QYX74F","created_at":"2026-07-05T11:18:21.677906+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":6,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.27147","citing_title":"How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance","ref_index":64,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27147","citing_title":"How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2604.27147","citing_title":"How to Guide Your Flow: Few-Step Alignment via Flow Map Reward Guidance","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19544","citing_title":"DT2IT-MRM: Debiased Preference Construction and Iterative Training for Multimodal Reward Modeling","ref_index":51,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13029","citing_title":"Visual Preference Optimization with Rubric Rewards","ref_index":34,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07872","citing_title":"Video Understanding Reward Modeling: A Robust Benchmark and Performant Reward Models","ref_index":33,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/U2QYX74FIHIC5AIK2BNY2NTVLP","json":"https://pith.science/pith/U2QYX74FIHIC5AIK2BNY2NTVLP.json","graph_json":"https://pith.science/api/pith-number/U2QYX74FIHIC5AIK2BNY2NTVLP/graph.json","events_json":"https://pith.science/api/pith-number/U2QYX74FIHIC5AIK2BNY2NTVLP/events.json","paper":"https://pith.science/paper/U2QYX74F"},"agent_actions":{"view_html":"https://pith.science/pith/U2QYX74FIHIC5AIK2BNY2NTVLP","download_json":"https://pith.science/pith/U2QYX74FIHIC5AIK2BNY2NTVLP.json","view_paper":"https://pith.science/paper/U2QYX74F","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.07263&json=true","fetch_graph":"https://pith.science/api/pith-number/U2QYX74FIHIC5AIK2BNY2NTVLP/graph.json","fetch_events":"https://pith.science/api/pith-number/U2QYX74FIHIC5AIK2BNY2NTVLP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/U2QYX74FIHIC5AIK2BNY2NTVLP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/U2QYX74FIHIC5AIK2BNY2NTVLP/action/storage_attestation","attest_author":"https://pith.science/pith/U2QYX74FIHIC5AIK2BNY2NTVLP/action/author_attestation","sign_citation":"https://pith.science/pith/U2QYX74FIHIC5AIK2BNY2NTVLP/action/citation_signature","submit_replication":"https://pith.science/pith/U2QYX74FIHIC5AIK2BNY2NTVLP/action/replication_record"}},"created_at":"2026-07-05T11:18:21.677906+00:00","updated_at":"2026-07-05T11:18:21.677906+00:00"}