{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:CFPRAWPFXMC46KS3DNT55DWW24","short_pith_number":"pith:CFPRAWPF","schema_version":"1.0","canonical_sha256":"115f1059e5bb05cf2a5b1b67de8ed6d718fe40f06196135e7ee20ed1f3ecebd1","source":{"kind":"arxiv","id":"2603.03143","version":2},"attestation_state":"computed","paper":{"title":"Edit in 2D, Verify in 3D: Reinforcement Learning for Multi-view Consistent Scene Editing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chunyu Lin, Guosheng Lin, Jiyuan Wang, Kang Liao, Lang Nie, Lei Sun, Xiangxiang Chu, Yunchao Wei, Yuyang Yin, Zhenlong Yuan, Zhi Cao","submitted_at":"2026-03-03T16:31:10Z","abstract_excerpt":"Leveraging the priors of 2D diffusion models for 3D editing has emerged as a promising paradigm. However, multi-view consistency remains challenging in edited results, and the extreme scarcity of paired 3D-consistent editing data makes supervised fine-tuning (SFT) impractical, despite its effectiveness for editing tasks. In this paper, we observe that, while generating multi-view consistent 3D content is highly challenging, verifying 3D consistency is tractable, naturally positioning reinforcement learning (RL) as a feasible solution. Motivated by this, we propose RL3DEdit, a single-pass frame"},"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":"2603.03143","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2026-03-03T16:31:10Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"22ce0da2e0f9140cd36270cb45377cd41c0eb310bc9df1e387e5eeb275ad5105","abstract_canon_sha256":"b47cad796ac203995d706ec7d628532b758c0063861f78d10e3003539d0ba732"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-06-30T01:17:35.142974Z","signature_b64":"W3cQreu3hqzfj97HvKil0OIJhZvcRoqW/6BB9lb8FCzbcgpg2TQB74ni2llAlDvtzqc6S0kdBiMkw202tM2qBA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"115f1059e5bb05cf2a5b1b67de8ed6d718fe40f06196135e7ee20ed1f3ecebd1","last_reissued_at":"2026-06-30T01:17:35.142223Z","signature_status":"signed_v1","first_computed_at":"2026-06-30T01:17:35.142223Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Edit in 2D, Verify in 3D: Reinforcement Learning for Multi-view Consistent Scene Editing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI"],"primary_cat":"cs.CV","authors_text":"Chunyu Lin, Guosheng Lin, Jiyuan Wang, Kang Liao, Lang Nie, Lei Sun, Xiangxiang Chu, Yunchao Wei, Yuyang Yin, Zhenlong Yuan, Zhi Cao","submitted_at":"2026-03-03T16:31:10Z","abstract_excerpt":"Leveraging the priors of 2D diffusion models for 3D editing has emerged as a promising paradigm. However, multi-view consistency remains challenging in edited results, and the extreme scarcity of paired 3D-consistent editing data makes supervised fine-tuning (SFT) impractical, despite its effectiveness for editing tasks. In this paper, we observe that, while generating multi-view consistent 3D content is highly challenging, verifying 3D consistency is tractable, naturally positioning reinforcement learning (RL) as a feasible solution. Motivated by this, we propose RL3DEdit, a single-pass frame"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2603.03143","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/2603.03143/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":"2603.03143","created_at":"2026-06-30T01:17:35.142321+00:00"},{"alias_kind":"arxiv_version","alias_value":"2603.03143v2","created_at":"2026-06-30T01:17:35.142321+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2603.03143","created_at":"2026-06-30T01:17:35.142321+00:00"},{"alias_kind":"pith_short_12","alias_value":"CFPRAWPFXMC4","created_at":"2026-06-30T01:17:35.142321+00:00"},{"alias_kind":"pith_short_16","alias_value":"CFPRAWPFXMC46KS3","created_at":"2026-06-30T01:17:35.142321+00:00"},{"alias_kind":"pith_short_8","alias_value":"CFPRAWPF","created_at":"2026-06-30T01:17:35.142321+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":5,"sample":[{"citing_arxiv_id":"2607.02220","citing_title":"DetailAnywhere: Fashion Detail Generation via Cross-Modal Feature Alignment Distillation","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2606.05142","citing_title":"GeM-NR: Geometry-Aware Multi-View Editing for Nonrigid Scene Changes","ref_index":57,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11723","citing_title":"CaC: Advancing Video Reward Models via Hierarchical Spatiotemporal Concentrating","ref_index":60,"is_internal_anchor":true},{"citing_arxiv_id":"2605.20682","citing_title":"IndusAgent: Reinforcing Open-Vocabulary Industrial Anomaly Detection with Agentic Tools","ref_index":87,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11723","citing_title":"CaC: Advancing Video Reward Models via Hierarchical Spatiotemporal Concentrating","ref_index":48,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/CFPRAWPFXMC46KS3DNT55DWW24","json":"https://pith.science/pith/CFPRAWPFXMC46KS3DNT55DWW24.json","graph_json":"https://pith.science/api/pith-number/CFPRAWPFXMC46KS3DNT55DWW24/graph.json","events_json":"https://pith.science/api/pith-number/CFPRAWPFXMC46KS3DNT55DWW24/events.json","paper":"https://pith.science/paper/CFPRAWPF"},"agent_actions":{"view_html":"https://pith.science/pith/CFPRAWPFXMC46KS3DNT55DWW24","download_json":"https://pith.science/pith/CFPRAWPFXMC46KS3DNT55DWW24.json","view_paper":"https://pith.science/paper/CFPRAWPF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2603.03143&json=true","fetch_graph":"https://pith.science/api/pith-number/CFPRAWPFXMC46KS3DNT55DWW24/graph.json","fetch_events":"https://pith.science/api/pith-number/CFPRAWPFXMC46KS3DNT55DWW24/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/CFPRAWPFXMC46KS3DNT55DWW24/action/timestamp_anchor","attest_storage":"https://pith.science/pith/CFPRAWPFXMC46KS3DNT55DWW24/action/storage_attestation","attest_author":"https://pith.science/pith/CFPRAWPFXMC46KS3DNT55DWW24/action/author_attestation","sign_citation":"https://pith.science/pith/CFPRAWPFXMC46KS3DNT55DWW24/action/citation_signature","submit_replication":"https://pith.science/pith/CFPRAWPFXMC46KS3DNT55DWW24/action/replication_record"}},"created_at":"2026-06-30T01:17:35.142321+00:00","updated_at":"2026-06-30T01:17:35.142321+00:00"}