{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:Z7NUCBYK7I573ZCKB47PFXXNRJ","short_pith_number":"pith:Z7NUCBYK","schema_version":"1.0","canonical_sha256":"cfdb41070afa3bfde44a0f3ef2deed8a67856a760c1f7a8111a10cd7a0f40f1f","source":{"kind":"arxiv","id":"2503.07065","version":1},"attestation_state":"computed","paper":{"title":"Boosting the Generalization and Reasoning of Vision Language Models with Curriculum Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ding Zou, Hongchen Luo, Huilin Deng, Rui Ma, Yang Cao, Yu Kang","submitted_at":"2025-03-10T08:48:50Z","abstract_excerpt":"While state-of-the-art vision-language models (VLMs) have demonstrated remarkable capabilities in complex visual-text tasks, their success heavily relies on massive model scaling, limiting their practical deployment. Small-scale VLMs offer a more practical alternative but face significant challenges when trained with traditional supervised fine-tuning (SFT), particularly in two aspects: out-of-domain (OOD) generalization and reasoning abilities, which significantly lags behind the contemporary Large language models (LLMs). To address these challenges, we propose Curriculum Reinforcement Finetu"},"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":"2503.07065","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-03-10T08:48:50Z","cross_cats_sorted":[],"title_canon_sha256":"5c81658ac165a5da5ebeb55c2aca68c2c51eb2cf0791712f511ffbe885e8a684","abstract_canon_sha256":"721a71b2d594fde70b18322eab9f2e398f6e89ebea2a421f797649616b885498"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:27:48.997552Z","signature_b64":"aXBCCloNgfTso8evDb4HV/8h3dPC32tcYxOTcESlhxb/gMcQnJK/kwkpsgs5HaVdH5KHBpW4ey6D9BdRqpE1AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"cfdb41070afa3bfde44a0f3ef2deed8a67856a760c1f7a8111a10cd7a0f40f1f","last_reissued_at":"2026-07-05T10:27:48.996954Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:27:48.996954Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Boosting the Generalization and Reasoning of Vision Language Models with Curriculum Reinforcement Learning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ding Zou, Hongchen Luo, Huilin Deng, Rui Ma, Yang Cao, Yu Kang","submitted_at":"2025-03-10T08:48:50Z","abstract_excerpt":"While state-of-the-art vision-language models (VLMs) have demonstrated remarkable capabilities in complex visual-text tasks, their success heavily relies on massive model scaling, limiting their practical deployment. Small-scale VLMs offer a more practical alternative but face significant challenges when trained with traditional supervised fine-tuning (SFT), particularly in two aspects: out-of-domain (OOD) generalization and reasoning abilities, which significantly lags behind the contemporary Large language models (LLMs). To address these challenges, we propose Curriculum Reinforcement Finetu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2503.07065","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/2503.07065/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":"2503.07065","created_at":"2026-07-05T10:27:48.997022+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.07065v1","created_at":"2026-07-05T10:27:48.997022+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.07065","created_at":"2026-07-05T10:27:48.997022+00:00"},{"alias_kind":"pith_short_12","alias_value":"Z7NUCBYK7I57","created_at":"2026-07-05T10:27:48.997022+00:00"},{"alias_kind":"pith_short_16","alias_value":"Z7NUCBYK7I573ZCK","created_at":"2026-07-05T10:27:48.997022+00:00"},{"alias_kind":"pith_short_8","alias_value":"Z7NUCBYK","created_at":"2026-07-05T10:27:48.997022+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":14,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.26196","citing_title":"From Structure to Synergy: A Survey of Vision-Language Perception Paradigm Evolution in Multimodal Large Language Models","ref_index":176,"is_internal_anchor":false},{"citing_arxiv_id":"2606.23557","citing_title":"Dense Reward for Multi-View 3D Reasoning with Global Maps and Local Views","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2606.09059","citing_title":"Stage-1 Controls the Entropy Regime, Not the Outcome","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06294","citing_title":"Towards One-to-Many Temporal Grounding","ref_index":46,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17037","citing_title":"D$^2$Evo: Dual Difficulty-Aware Self-Evolution for Data-Efficient Reinforcement Learning","ref_index":56,"is_internal_anchor":false},{"citing_arxiv_id":"2507.00748","citing_title":"Improving the Reasoning of Multi-Image Grounding in MLLMs via Reinforcement Learning","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2504.06958","citing_title":"VideoChat-R1: Enhancing Spatio-Temporal Perception via Reinforcement Fine-Tuning","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2602.19974","citing_title":"RL-RIG: A Generative Spatial Reasoner via Intrinsic Reflection","ref_index":6,"is_internal_anchor":false},{"citing_arxiv_id":"2503.12605","citing_title":"Multimodal Chain-of-Thought Reasoning: A Comprehensive Survey","ref_index":245,"is_internal_anchor":false},{"citing_arxiv_id":"2504.07615","citing_title":"VLM-R1: A Stable and Generalizable R1-style Large Vision-Language Model","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10761","citing_title":"RadThinking: A Dataset for Longitudinal Clinical Reasoning in Radiology","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.09188","citing_title":"DARE: Difficulty-Adaptive Reinforcement Learning with Co-Evolved Difficulty Estimation","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05341","citing_title":"Curr-RLCER:Curriculum Reinforcement Learning For Coherence Explainable Recommendation","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2604.16557","citing_title":"S-GRPO: Unified Post-Training for Large Vision-Language Models","ref_index":12,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/Z7NUCBYK7I573ZCKB47PFXXNRJ","json":"https://pith.science/pith/Z7NUCBYK7I573ZCKB47PFXXNRJ.json","graph_json":"https://pith.science/api/pith-number/Z7NUCBYK7I573ZCKB47PFXXNRJ/graph.json","events_json":"https://pith.science/api/pith-number/Z7NUCBYK7I573ZCKB47PFXXNRJ/events.json","paper":"https://pith.science/paper/Z7NUCBYK"},"agent_actions":{"view_html":"https://pith.science/pith/Z7NUCBYK7I573ZCKB47PFXXNRJ","download_json":"https://pith.science/pith/Z7NUCBYK7I573ZCKB47PFXXNRJ.json","view_paper":"https://pith.science/paper/Z7NUCBYK","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.07065&json=true","fetch_graph":"https://pith.science/api/pith-number/Z7NUCBYK7I573ZCKB47PFXXNRJ/graph.json","fetch_events":"https://pith.science/api/pith-number/Z7NUCBYK7I573ZCKB47PFXXNRJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/Z7NUCBYK7I573ZCKB47PFXXNRJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/Z7NUCBYK7I573ZCKB47PFXXNRJ/action/storage_attestation","attest_author":"https://pith.science/pith/Z7NUCBYK7I573ZCKB47PFXXNRJ/action/author_attestation","sign_citation":"https://pith.science/pith/Z7NUCBYK7I573ZCKB47PFXXNRJ/action/citation_signature","submit_replication":"https://pith.science/pith/Z7NUCBYK7I573ZCKB47PFXXNRJ/action/replication_record"}},"created_at":"2026-07-05T10:27:48.997022+00:00","updated_at":"2026-07-05T10:27:48.997022+00:00"}