{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:R5UDKTPTYBNYTROJBECV45AWHU","short_pith_number":"pith:R5UDKTPT","schema_version":"1.0","canonical_sha256":"8f68354df3c05b89c5c909055e74163d13b426e7b174e1b1bfa301de942bfcc7","source":{"kind":"arxiv","id":"2505.13031","version":2},"attestation_state":"computed","paper":{"title":"MindOmni: Unleashing Reasoning Generation in Vision Language Models with RGPO","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Lin Song, Wei Huang, Xiaojuan Qi, Xiu Li, Yicheng Xiao, Yingmin Luo, Ying Shan, Yukang Chen, Yukang Gan, Yuxin Chen","submitted_at":"2025-05-19T12:17:04Z","abstract_excerpt":"Recent text-to-image systems face limitations in handling multimodal inputs and complex reasoning tasks. We introduce MindOmni, a unified multimodal large language model that addresses these challenges by incorporating reasoning generation through reinforcement learning. MindOmni leverages a three-phase training strategy: i) design of a unified vision language model with a decoder-only diffusion module, ii) supervised fine-tuning with Chain-of-Thought (CoT) instruction data, and iii) our proposed Reasoning Generation Policy Optimization (RGPO) algorithm, utilizing multimodal feedback to effect"},"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":"2505.13031","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.AI","submitted_at":"2025-05-19T12:17:04Z","cross_cats_sorted":[],"title_canon_sha256":"834f68cb41708af5f657b8d262e22b72e4726accd5101fa638a74078a2e77a91","abstract_canon_sha256":"3aa0e48c178080202e6972ce983d68888a66c7b7828fff17c4859754aaa0ca47"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:19:52.743967Z","signature_b64":"sSM4DqjYNE3OSBBhOy5m/Tmoht3hHUBzndOqNU+MX7xdz+xUQf/B1nZH+zAC+Y6qpFfE9GyXBXw0ieelOx/eDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8f68354df3c05b89c5c909055e74163d13b426e7b174e1b1bfa301de942bfcc7","last_reissued_at":"2026-07-05T11:19:52.743459Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:19:52.743459Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MindOmni: Unleashing Reasoning Generation in Vision Language Models with RGPO","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.AI","authors_text":"Lin Song, Wei Huang, Xiaojuan Qi, Xiu Li, Yicheng Xiao, Yingmin Luo, Ying Shan, Yukang Chen, Yukang Gan, Yuxin Chen","submitted_at":"2025-05-19T12:17:04Z","abstract_excerpt":"Recent text-to-image systems face limitations in handling multimodal inputs and complex reasoning tasks. We introduce MindOmni, a unified multimodal large language model that addresses these challenges by incorporating reasoning generation through reinforcement learning. MindOmni leverages a three-phase training strategy: i) design of a unified vision language model with a decoder-only diffusion module, ii) supervised fine-tuning with Chain-of-Thought (CoT) instruction data, and iii) our proposed Reasoning Generation Policy Optimization (RGPO) algorithm, utilizing multimodal feedback to effect"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.13031","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/2505.13031/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":"2505.13031","created_at":"2026-07-05T11:19:52.743517+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.13031v2","created_at":"2026-07-05T11:19:52.743517+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.13031","created_at":"2026-07-05T11:19:52.743517+00:00"},{"alias_kind":"pith_short_12","alias_value":"R5UDKTPTYBNY","created_at":"2026-07-05T11:19:52.743517+00:00"},{"alias_kind":"pith_short_16","alias_value":"R5UDKTPTYBNYTROJ","created_at":"2026-07-05T11:19:52.743517+00:00"},{"alias_kind":"pith_short_8","alias_value":"R5UDKTPT","created_at":"2026-07-05T11:19:52.743517+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.04264","citing_title":"UniCanvas: A Diffusion-base Unified Model for Text-in-Image Joint Generation","ref_index":57,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04128","citing_title":"JoyAI-Image: Awaking Spatial Intelligence in Unified Multimodal Understanding and Generation","ref_index":91,"is_internal_anchor":false},{"citing_arxiv_id":"2605.07334","citing_title":"RCoT-Seg: Reinforced Chain-of-Thought for Video Reasoning and Segmentation","ref_index":28,"is_internal_anchor":false},{"citing_arxiv_id":"2604.04911","citing_title":"SpatialEdit: Benchmarking Fine-Grained Image Spatial Editing","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2605.04128","citing_title":"JoyAI-Image: Awaking Spatial Intelligence in Unified Multimodal Understanding and Generation","ref_index":91,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/R5UDKTPTYBNYTROJBECV45AWHU","json":"https://pith.science/pith/R5UDKTPTYBNYTROJBECV45AWHU.json","graph_json":"https://pith.science/api/pith-number/R5UDKTPTYBNYTROJBECV45AWHU/graph.json","events_json":"https://pith.science/api/pith-number/R5UDKTPTYBNYTROJBECV45AWHU/events.json","paper":"https://pith.science/paper/R5UDKTPT"},"agent_actions":{"view_html":"https://pith.science/pith/R5UDKTPTYBNYTROJBECV45AWHU","download_json":"https://pith.science/pith/R5UDKTPTYBNYTROJBECV45AWHU.json","view_paper":"https://pith.science/paper/R5UDKTPT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.13031&json=true","fetch_graph":"https://pith.science/api/pith-number/R5UDKTPTYBNYTROJBECV45AWHU/graph.json","fetch_events":"https://pith.science/api/pith-number/R5UDKTPTYBNYTROJBECV45AWHU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/R5UDKTPTYBNYTROJBECV45AWHU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/R5UDKTPTYBNYTROJBECV45AWHU/action/storage_attestation","attest_author":"https://pith.science/pith/R5UDKTPTYBNYTROJBECV45AWHU/action/author_attestation","sign_citation":"https://pith.science/pith/R5UDKTPTYBNYTROJBECV45AWHU/action/citation_signature","submit_replication":"https://pith.science/pith/R5UDKTPTYBNYTROJBECV45AWHU/action/replication_record"}},"created_at":"2026-07-05T11:19:52.743517+00:00","updated_at":"2026-07-05T11:19:52.743517+00:00"}