{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:ZAKNLGD2YQFUOJIB7UI32TXFHU","short_pith_number":"pith:ZAKNLGD2","schema_version":"1.0","canonical_sha256":"c814d5987ac40b472501fd11bd4ee53d07720c7ffd55fad8eb835aa74985969e","source":{"kind":"arxiv","id":"2505.23380","version":1},"attestation_state":"computed","paper":{"title":"UniRL: Self-Improving Unified Multimodal Models via Supervised and Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Mike Zheng Shou, Weijia Mao, Zhenheng Yang","submitted_at":"2025-05-29T12:00:15Z","abstract_excerpt":"Unified multimodal large language models such as Show-o and Janus have achieved strong performance across both generation and understanding tasks. However, these models typically rely on large-scale datasets and require substantial computation during the pretraining stage. In addition, several post-training methods have been proposed, but they often depend on external data or are limited to task-specific customization. In this work, we introduce UniRL, a self-improving post-training approach. Our approach enables the model to generate images from prompts and use them as training data in each i"},"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":"2505.23380","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-05-29T12:00:15Z","cross_cats_sorted":[],"title_canon_sha256":"389e1b618202f02a9ad9a762b13870f78e028e99392c22bbd77d11e0560a3bc0","abstract_canon_sha256":"b52288b12ce73c1590e97b9b2324266c60e0ce0d120c900ba417d1c413547c4a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:12:00.726648Z","signature_b64":"+EnWciPy6ZD2HI9zzd662cdycTN83M8TScZH5pszJRtKQg3QQ0Q1yL7sD+eOyDOM3knG11d8hrhzoCa9Sil/Aw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c814d5987ac40b472501fd11bd4ee53d07720c7ffd55fad8eb835aa74985969e","last_reissued_at":"2026-07-05T11:12:00.726065Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:12:00.726065Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"UniRL: Self-Improving Unified Multimodal Models via Supervised and Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Mike Zheng Shou, Weijia Mao, Zhenheng Yang","submitted_at":"2025-05-29T12:00:15Z","abstract_excerpt":"Unified multimodal large language models such as Show-o and Janus have achieved strong performance across both generation and understanding tasks. However, these models typically rely on large-scale datasets and require substantial computation during the pretraining stage. In addition, several post-training methods have been proposed, but they often depend on external data or are limited to task-specific customization. In this work, we introduce UniRL, a self-improving post-training approach. Our approach enables the model to generate images from prompts and use them as training data in each i"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2505.23380","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/2505.23380/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":"2505.23380","created_at":"2026-07-05T11:12:00.726123+00:00"},{"alias_kind":"arxiv_version","alias_value":"2505.23380v1","created_at":"2026-07-05T11:12:00.726123+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2505.23380","created_at":"2026-07-05T11:12:00.726123+00:00"},{"alias_kind":"pith_short_12","alias_value":"ZAKNLGD2YQFU","created_at":"2026-07-05T11:12:00.726123+00:00"},{"alias_kind":"pith_short_16","alias_value":"ZAKNLGD2YQFUOJIB","created_at":"2026-07-05T11:12:00.726123+00:00"},{"alias_kind":"pith_short_8","alias_value":"ZAKNLGD2","created_at":"2026-07-05T11:12:00.726123+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.27376","citing_title":"Ask, Solve, Generate: Self-Evolving Unified Multimodal Understanding and Generation via Self-Consistency Rewards","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25343","citing_title":"Toward Native Multimodal Modeling: A Roadmap","ref_index":195,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17766","citing_title":"LatentUMM: Dual Latent Alignment for Unified Multimodal Models","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/ZAKNLGD2YQFUOJIB7UI32TXFHU","json":"https://pith.science/pith/ZAKNLGD2YQFUOJIB7UI32TXFHU.json","graph_json":"https://pith.science/api/pith-number/ZAKNLGD2YQFUOJIB7UI32TXFHU/graph.json","events_json":"https://pith.science/api/pith-number/ZAKNLGD2YQFUOJIB7UI32TXFHU/events.json","paper":"https://pith.science/paper/ZAKNLGD2"},"agent_actions":{"view_html":"https://pith.science/pith/ZAKNLGD2YQFUOJIB7UI32TXFHU","download_json":"https://pith.science/pith/ZAKNLGD2YQFUOJIB7UI32TXFHU.json","view_paper":"https://pith.science/paper/ZAKNLGD2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2505.23380&json=true","fetch_graph":"https://pith.science/api/pith-number/ZAKNLGD2YQFUOJIB7UI32TXFHU/graph.json","fetch_events":"https://pith.science/api/pith-number/ZAKNLGD2YQFUOJIB7UI32TXFHU/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/ZAKNLGD2YQFUOJIB7UI32TXFHU/action/timestamp_anchor","attest_storage":"https://pith.science/pith/ZAKNLGD2YQFUOJIB7UI32TXFHU/action/storage_attestation","attest_author":"https://pith.science/pith/ZAKNLGD2YQFUOJIB7UI32TXFHU/action/author_attestation","sign_citation":"https://pith.science/pith/ZAKNLGD2YQFUOJIB7UI32TXFHU/action/citation_signature","submit_replication":"https://pith.science/pith/ZAKNLGD2YQFUOJIB7UI32TXFHU/action/replication_record"}},"created_at":"2026-07-05T11:12:00.726123+00:00","updated_at":"2026-07-05T11:12:00.726123+00:00"}