{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:MLVB2TIWDNTVNEEDGGKZ4NODID","short_pith_number":"pith:MLVB2TIW","schema_version":"1.0","canonical_sha256":"62ea1d4d161b6756908331959e35c340ceec756c31d372ce554f1d8c6a15f9f1","source":{"kind":"arxiv","id":"2502.06812","version":2},"attestation_state":"computed","paper":{"title":"Harness Local Rewards for Global Benefits: Effective Text-to-Video Generation Alignment with Patch-level Reward Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.GR"],"primary_cat":"cs.LG","authors_text":"Chenyan Xiong, Haihong Tang, Shuting Wang, Zhicheng Dou","submitted_at":"2025-02-04T21:10:25Z","abstract_excerpt":"The emergence of diffusion models (DMs) has significantly improved the quality of text-to-video generation models (VGMs). However, current VGM optimization primarily emphasizes the global quality of videos, overlooking localized errors, which leads to suboptimal generation capabilities. To address this issue, we propose a post-training strategy for VGMs, HALO, which explicitly incorporates local feedback from a patch reward model, providing detailed and comprehensive training signals with the video reward model for advanced VGM optimization. To develop an effective patch reward model, we disti"},"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":"2502.06812","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2025-02-04T21:10:25Z","cross_cats_sorted":["cs.GR"],"title_canon_sha256":"1eed06ee97afd2f55962961ad1b8bfd2e8ee9af3a2e294faee4c6a5adc8f56ca","abstract_canon_sha256":"64d705452664cb58c884a2b1f5b213b735f7510f88abfb685b076adaf677e32c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:15:48.975940Z","signature_b64":"DYu76IeCtmvyBmSHw3gDiUJcklGDHJzk/snef3jXJsAT1290qYEFGfk6O0kVX8uF+xRkIdkOZWv9J4YqbX9WBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"62ea1d4d161b6756908331959e35c340ceec756c31d372ce554f1d8c6a15f9f1","last_reissued_at":"2026-07-05T10:15:48.975395Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:15:48.975395Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Harness Local Rewards for Global Benefits: Effective Text-to-Video Generation Alignment with Patch-level Reward Models","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.GR"],"primary_cat":"cs.LG","authors_text":"Chenyan Xiong, Haihong Tang, Shuting Wang, Zhicheng Dou","submitted_at":"2025-02-04T21:10:25Z","abstract_excerpt":"The emergence of diffusion models (DMs) has significantly improved the quality of text-to-video generation models (VGMs). However, current VGM optimization primarily emphasizes the global quality of videos, overlooking localized errors, which leads to suboptimal generation capabilities. To address this issue, we propose a post-training strategy for VGMs, HALO, which explicitly incorporates local feedback from a patch reward model, providing detailed and comprehensive training signals with the video reward model for advanced VGM optimization. To develop an effective patch reward model, we disti"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.06812","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/2502.06812/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":"2502.06812","created_at":"2026-07-05T10:15:48.975456+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.06812v2","created_at":"2026-07-05T10:15:48.975456+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.06812","created_at":"2026-07-05T10:15:48.975456+00:00"},{"alias_kind":"pith_short_12","alias_value":"MLVB2TIWDNTV","created_at":"2026-07-05T10:15:48.975456+00:00"},{"alias_kind":"pith_short_16","alias_value":"MLVB2TIWDNTVNEED","created_at":"2026-07-05T10:15:48.975456+00:00"},{"alias_kind":"pith_short_8","alias_value":"MLVB2TIW","created_at":"2026-07-05T10:15:48.975456+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.25759","citing_title":"Towards Anatomically Plausible Human Image Generation via Synthetic Localized Preferences","ref_index":38,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/MLVB2TIWDNTVNEEDGGKZ4NODID","json":"https://pith.science/pith/MLVB2TIWDNTVNEEDGGKZ4NODID.json","graph_json":"https://pith.science/api/pith-number/MLVB2TIWDNTVNEEDGGKZ4NODID/graph.json","events_json":"https://pith.science/api/pith-number/MLVB2TIWDNTVNEEDGGKZ4NODID/events.json","paper":"https://pith.science/paper/MLVB2TIW"},"agent_actions":{"view_html":"https://pith.science/pith/MLVB2TIWDNTVNEEDGGKZ4NODID","download_json":"https://pith.science/pith/MLVB2TIWDNTVNEEDGGKZ4NODID.json","view_paper":"https://pith.science/paper/MLVB2TIW","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.06812&json=true","fetch_graph":"https://pith.science/api/pith-number/MLVB2TIWDNTVNEEDGGKZ4NODID/graph.json","fetch_events":"https://pith.science/api/pith-number/MLVB2TIWDNTVNEEDGGKZ4NODID/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/MLVB2TIWDNTVNEEDGGKZ4NODID/action/timestamp_anchor","attest_storage":"https://pith.science/pith/MLVB2TIWDNTVNEEDGGKZ4NODID/action/storage_attestation","attest_author":"https://pith.science/pith/MLVB2TIWDNTVNEEDGGKZ4NODID/action/author_attestation","sign_citation":"https://pith.science/pith/MLVB2TIWDNTVNEEDGGKZ4NODID/action/citation_signature","submit_replication":"https://pith.science/pith/MLVB2TIWDNTVNEEDGGKZ4NODID/action/replication_record"}},"created_at":"2026-07-05T10:15:48.975456+00:00","updated_at":"2026-07-05T10:15:48.975456+00:00"}