{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:SGUFOMNHKQL4IFMWYUVCLHHQH6","short_pith_number":"pith:SGUFOMNH","schema_version":"1.0","canonical_sha256":"91a85731a75417c41596c52a259cf03f90f2493edb555761d42ec860aaa2051f","source":{"kind":"arxiv","id":"2503.07536","version":2},"attestation_state":"computed","paper":{"title":"LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RL","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Strengthening reasoning first on text data then transferring to images improves 3B multimodal models without extra multimodal data.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Gongrui Zhang, Jie Liu, Kai Yang, Miaosen Zhang, Qipeng Zhu, Xin Geng, Xingzhong Xu, Xu Yang, Yingzhe Peng, Zhiyuan You","submitted_at":"2025-03-10T17:04:14Z","abstract_excerpt":"Enhancing reasoning in Large Multimodal Models (LMMs) faces unique challenges from the complex interplay between visual perception and logical reasoning, particularly in compact 3B-parameter architectures where architectural constraints limit reasoning capacity and modality alignment.\n  While rule-based reinforcement learning (RL) excels in text-only domains, its multimodal extension confronts two critical barriers: (1) data limitations due to ambiguous answers and scarce complex reasoning examples, and (2) degraded foundational reasoning induced by multimodal pretraining. To address these cha"},"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":true,"formal_links_present":true},"canonical_record":{"source":{"id":"2503.07536","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CL","submitted_at":"2025-03-10T17:04:14Z","cross_cats_sorted":["cs.AI"],"title_canon_sha256":"43e5c65607dd7bda3cdd188b43717b8d938f768ed61b7ba0f73373e48bba2933","abstract_canon_sha256":"153a64ed5dc6d65cf29c723c8da3f7bac130de117b7afe63774316ea84075c24"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:47.472006Z","signature_b64":"nuf+6QGYCQHN99qO4zCtQu71ndGx8bfizMXlsVpRIU9b522Hv1M6ERd9kFH5/ynZUtX+shKNeKTrWicpwmahDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"91a85731a75417c41596c52a259cf03f90f2493edb555761d42ec860aaa2051f","last_reissued_at":"2026-05-17T23:38:47.471538Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:47.471538Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RL","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Strengthening reasoning first on text data then transferring to images improves 3B multimodal models without extra multimodal data.","cross_cats":["cs.AI"],"primary_cat":"cs.CL","authors_text":"Gongrui Zhang, Jie Liu, Kai Yang, Miaosen Zhang, Qipeng Zhu, Xin Geng, Xingzhong Xu, Xu Yang, Yingzhe Peng, Zhiyuan You","submitted_at":"2025-03-10T17:04:14Z","abstract_excerpt":"Enhancing reasoning in Large Multimodal Models (LMMs) faces unique challenges from the complex interplay between visual perception and logical reasoning, particularly in compact 3B-parameter architectures where architectural constraints limit reasoning capacity and modality alignment.\n  While rule-based reinforcement learning (RL) excels in text-only domains, its multimodal extension confronts two critical barriers: (1) data limitations due to ambiguous answers and scarce complex reasoning examples, and (2) degraded foundational reasoning induced by multimodal pretraining. To address these cha"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"text-based reasoning enhancement enables effective multimodal generalization, offering a data-efficient paradigm that bypasses costly high-quality multimodal training data.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That reasoning skills strengthened via text-only rule-based RL transfer to multimodal inputs without substantial interference from visual perception components or degradation of the foundational reasoning.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A two-stage RL framework first boosts text reasoning in 3B LMMs then adapts it to multimodal inputs, producing modest benchmark gains of 4.5-4.8%.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Strengthening reasoning first on text data then transferring to images improves 3B multimodal models without extra multimodal data.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"86c3f6d99d27df1b124bf81641972e85a603eca4101442a220dc3281f3f85af3"},"source":{"id":"2503.07536","kind":"arxiv","version":2},"verdict":{"id":"9c7ebaba-1a3d-46f4-a4f2-4d7f1402af82","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-16T15:11:35.828514Z","strongest_claim":"text-based reasoning enhancement enables effective multimodal generalization, offering a data-efficient paradigm that bypasses costly high-quality multimodal training data.","one_line_summary":"A two-stage RL framework first boosts text reasoning in 3B LMMs then adapts it to multimodal inputs, producing modest benchmark gains of 4.5-4.8%.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That reasoning skills strengthened via text-only rule-based RL transfer to multimodal inputs without substantial interference from visual perception components or degradation of the foundational reasoning.","pith_extraction_headline":"Strengthening reasoning first on text data then transferring to images improves 3B multimodal models without extra multimodal data."},"references":{"count":117,"sample":[{"doi":"","year":null,"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","ref_index":1,"cited_arxiv_id":"2303.08774","is_internal_anchor":true},{"doi":"","year":2015,"title":"Lawrence Zitnick, Devi Parikh, and Dhruv Ba- tra","work_id":"7945da46-3915-4108-853a-ab77b7e0cd3c","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":null,"title":"Flamingo: A visual language model for few-shot learning","work_id":"1863a034-835c-4304-b6c6-32734dcf9e1a","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Qwen Technical Report","work_id":"bb1fd52f-6b2f-437c-9516-37bdf6eb9be8","ref_index":4,"cited_arxiv_id":"2309.16609","is_internal_anchor":true},{"doi":"","year":2023,"title":"Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond","work_id":"cbc2bb21-b6bb-46c0-80bf-107e195ffe10","ref_index":5,"cited_arxiv_id":"2308.12966","is_internal_anchor":true}],"resolved_work":117,"snapshot_sha256":"19d1ebaaa5411c244a51a482d94aacad3b045979c3b0e1835cfd65922d00882d","internal_anchors":31},"formal_canon":{"evidence_count":2,"snapshot_sha256":"1f56fb3c6ceb513d5107c257c4adc2804883b4df9c9db298f66864c895bfa6d4"},"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":"2503.07536","created_at":"2026-05-17T23:38:47.471610+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.07536v2","created_at":"2026-05-17T23:38:47.471610+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.07536","created_at":"2026-05-17T23:38:47.471610+00:00"},{"alias_kind":"pith_short_12","alias_value":"SGUFOMNHKQL4","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"SGUFOMNHKQL4IFMW","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"SGUFOMNH","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":35,"internal_anchor_count":35,"sample":[{"citing_arxiv_id":"2605.23281","citing_title":"DepthAgent: Towards Better Universal Depth Estimation via Sample-wise Expert Selection","ref_index":47,"is_internal_anchor":true},{"citing_arxiv_id":"2605.20342","citing_title":"ParaVT: Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2511.14159","citing_title":"MVI-Bench: A Comprehensive Benchmark for Evaluating Robustness to Misleading Visual Inputs in LVLMs","ref_index":39,"is_internal_anchor":true},{"citing_arxiv_id":"2605.20342","citing_title":"ParaVT: Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2605.20682","citing_title":"IndusAgent: Reinforcing Open-Vocabulary Industrial Anomaly Detection with Agentic Tools","ref_index":65,"is_internal_anchor":true},{"citing_arxiv_id":"2605.14068","citing_title":"CurveBench: A Benchmark for Exact Topological Reasoning over Nested Jordan Curves","ref_index":40,"is_internal_anchor":true},{"citing_arxiv_id":"2605.20177","citing_title":"From Seeing to Thinking: Decoupling Perception and Reasoning Improves Post-Training of Vision-Language Models","ref_index":20,"is_internal_anchor":true},{"citing_arxiv_id":"2509.02547","citing_title":"The Landscape of Agentic Reinforcement Learning for LLMs: A Survey","ref_index":227,"is_internal_anchor":true},{"citing_arxiv_id":"2509.21976","citing_title":"Geo-R1: Improving Few-Shot Geospatial Referring Expression Understanding with Reinforcement Fine-Tuning","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2509.22746","citing_title":"Mixture-of-Visual-Thoughts: Exploring Context-Adaptive Reasoning Mode Selection for General Visual Reasoning","ref_index":32,"is_internal_anchor":true},{"citing_arxiv_id":"2509.23322","citing_title":"Mitigating Visual Context Degradation in Large Multimodal Models: A Training-Free Decoupled Agentic Framework","ref_index":23,"is_internal_anchor":true},{"citing_arxiv_id":"2510.21122","citing_title":"NoisyGRPO: Incentivizing Multimodal CoT Reasoning via Noise Injection and Bayesian Estimation","ref_index":38,"is_internal_anchor":true},{"citing_arxiv_id":"2511.20814","citing_title":"SPHINX: A Synthetic Environment for Visual Perception and Reasoning","ref_index":39,"is_internal_anchor":true},{"citing_arxiv_id":"2512.16918","citing_title":"AdaTooler-V: Adaptive Tool-Use for Images and Videos","ref_index":47,"is_internal_anchor":true},{"citing_arxiv_id":"2601.04442","citing_title":"Addressing Overthinking in Large Vision-Language Models via Gated Perception-Reasoning Optimization","ref_index":9,"is_internal_anchor":true},{"citing_arxiv_id":"2503.12937","citing_title":"R1-VL: Learning to Reason with Multimodal Large Language Models via Step-wise Group Relative Policy Optimization","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2601.06993","citing_title":"Can Textual Reasoning Improve the Performance of MLLMs on Fine-grained Visual Classification?","ref_index":26,"is_internal_anchor":true},{"citing_arxiv_id":"2504.06958","citing_title":"VideoChat-R1: Enhancing Spatio-Temporal Perception via Reinforcement Fine-Tuning","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2509.24251","citing_title":"Latent Visual Reasoning","ref_index":18,"is_internal_anchor":true},{"citing_arxiv_id":"2504.01805","citing_title":"SpaceR: Reinforcing MLLMs in Video Spatial Reasoning","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2605.14068","citing_title":"CurveBench: A Benchmark for Exact Topological Reasoning over Nested Jordan Curves","ref_index":40,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12163","citing_title":"Self-Consistent Latent Reasoning: Long Latent Sequence Reasoning for Vision-Language Model","ref_index":30,"is_internal_anchor":true},{"citing_arxiv_id":"2603.27494","citing_title":"Learning to Focus and Precise Cropping: A Reinforcement Learning Framework with Information Gaps and Grounding Loss for MLLMs","ref_index":29,"is_internal_anchor":true},{"citing_arxiv_id":"2604.03307","citing_title":"V-Reflection: Transforming MLLMs from Passive Observers to Active Interrogators","ref_index":16,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12163","citing_title":"Self-Consistent Latent Reasoning: Long Latent Sequence Reasoning for Vision-Language Model","ref_index":30,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6","json":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6.json","graph_json":"https://pith.science/api/pith-number/SGUFOMNHKQL4IFMWYUVCLHHQH6/graph.json","events_json":"https://pith.science/api/pith-number/SGUFOMNHKQL4IFMWYUVCLHHQH6/events.json","paper":"https://pith.science/paper/SGUFOMNH"},"agent_actions":{"view_html":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6","download_json":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6.json","view_paper":"https://pith.science/paper/SGUFOMNH","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.07536&json=true","fetch_graph":"https://pith.science/api/pith-number/SGUFOMNHKQL4IFMWYUVCLHHQH6/graph.json","fetch_events":"https://pith.science/api/pith-number/SGUFOMNHKQL4IFMWYUVCLHHQH6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/action/storage_attestation","attest_author":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/action/author_attestation","sign_citation":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/action/citation_signature","submit_replication":"https://pith.science/pith/SGUFOMNHKQL4IFMWYUVCLHHQH6/action/replication_record"}},"created_at":"2026-05-17T23:38:47.471610+00:00","updated_at":"2026-05-17T23:38:47.471610+00:00"}