{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:AWCZRZQ4QCF2K4O2MUR2KYFJ4P","short_pith_number":"pith:AWCZRZQ4","schema_version":"1.0","canonical_sha256":"058598e61c808ba571da6523a560a9e3c7a30acc1708721a83ab8864b033787e","source":{"kind":"arxiv","id":"2503.05236","version":2},"attestation_state":"computed","paper":{"title":"Unified Reward Model for Multimodal Understanding and Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A single reward model trained jointly on image and video tasks improves preference alignment for both understanding and generation.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cheng Jin, Hao Li, Jiaqi Wang, Yibin Wang, Yuhang Zang","submitted_at":"2025-03-07T08:36:05Z","abstract_excerpt":"Recent advances in human preference alignment have significantly improved multimodal generation and understanding. A key approach is to train reward models that provide supervision signals for preference optimization. However, existing reward models are often task-specific, limiting their adaptability across diverse visual applications. We also argue that a reward model that jointly learning to assess multiple vision tasks may foster a synergistic effect, where improved image understanding enhances image generation assessment, and refined image evaluation benefits video assessment through bett"},"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":"2503.05236","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2025-03-07T08:36:05Z","cross_cats_sorted":[],"title_canon_sha256":"2aedac95cbbedd1322221f81bf6fbba2b5fd14fda6c12942a05f73b48664f4f8","abstract_canon_sha256":"61deb671ab72436b4fe9934b639cef553965b827b86a5f6c7be13ed84a2aa7c3"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T03:15:18.211161Z","signature_b64":"/9h7O5VyX2IpkoNvNl3rlAAqk03k8MHcWeQzQx+FU3MYSp6aPg/iiFeoibjsCw1yYlkrS01+HNw96Be7lIPDBw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"058598e61c808ba571da6523a560a9e3c7a30acc1708721a83ab8864b033787e","last_reissued_at":"2026-05-18T03:15:18.210251Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T03:15:18.210251Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unified Reward Model for Multimodal Understanding and Generation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"A single reward model trained jointly on image and video tasks improves preference alignment for both understanding and generation.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Cheng Jin, Hao Li, Jiaqi Wang, Yibin Wang, Yuhang Zang","submitted_at":"2025-03-07T08:36:05Z","abstract_excerpt":"Recent advances in human preference alignment have significantly improved multimodal generation and understanding. A key approach is to train reward models that provide supervision signals for preference optimization. However, existing reward models are often task-specific, limiting their adaptability across diverse visual applications. We also argue that a reward model that jointly learning to assess multiple vision tasks may foster a synergistic effect, where improved image understanding enhances image generation assessment, and refined image evaluation benefits video assessment through bett"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"jointly learning to assess diverse visual tasks yields substantial mutual benefits... achieving consistent improvements across each domain.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The large-scale human preference dataset accurately represents human judgments across tasks and the two-stage filtering strategy produces high-quality, unbiased preference pairs without introducing selection artifacts.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"UnifiedReward is the first unified reward model that jointly assesses multimodal understanding and generation to provide better preference signals for aligning vision models via DPO.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"A single reward model trained jointly on image and video tasks improves preference alignment for both understanding and generation.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"a04782d1d631e82f96d7da719f6f7958ce15ad318815067e1e7a0c81fc365900"},"source":{"id":"2503.05236","kind":"arxiv","version":2},"verdict":{"id":"27542910-c752-4be9-9a1f-1c30da435b53","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-14T00:39:50.045426Z","strongest_claim":"jointly learning to assess diverse visual tasks yields substantial mutual benefits... achieving consistent improvements across each domain.","one_line_summary":"UnifiedReward is the first unified reward model that jointly assesses multimodal understanding and generation to provide better preference signals for aligning vision models via DPO.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The large-scale human preference dataset accurately represents human judgments across tasks and the two-stage filtering strategy produces high-quality, unbiased preference pairs without introducing selection artifacts.","pith_extraction_headline":"A single reward model trained jointly on image and video tasks improves preference alignment for both understanding and generation."},"references":{"count":64,"sample":[{"doi":"","year":2024,"title":"Diffusion model alignment using direct preference optimization","work_id":"3bbf25f4-3cec-4668-a9d1-8ddccee8afd8","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Videodpo: Omni-preference alignment for video diffusion generation","work_id":"47508379-7568-448f-b406-1b4e9bbb181b","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Lift: Leveraging human feedback for text-to-video model alignment","work_id":"ab339bba-4351-45db-841d-de086f9c0d67","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"Llava-critic: Learning to evaluate multimodal models","work_id":"2d8a9688-7f2b-4224-8107-31c7945ce2cd","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Internlm-xcomposer2.5-reward: A simple yet effective multi-modal reward model","work_id":"499855fc-7d44-43d3-ab3a-1038d2ccd0bc","ref_index":6,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":64,"snapshot_sha256":"651a51160f69de7cdb1a714c615a241622a86798965a893fba208eafa2cf1838","internal_anchors":21},"formal_canon":{"evidence_count":2,"snapshot_sha256":"bc8e6693c3ba327a9134bd49e0e50656b6f9543d25273089ee312b7696bb487f"},"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":"2503.05236","created_at":"2026-05-18T03:15:18.210440+00:00"},{"alias_kind":"arxiv_version","alias_value":"2503.05236v2","created_at":"2026-05-18T03:15:18.210440+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2503.05236","created_at":"2026-05-18T03:15:18.210440+00:00"},{"alias_kind":"pith_short_12","alias_value":"AWCZRZQ4QCF2","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"AWCZRZQ4QCF2K4O2","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"AWCZRZQ4","created_at":"2026-05-18T12:33:37.589309+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":26,"internal_anchor_count":26,"sample":[{"citing_arxiv_id":"2512.04678","citing_title":"Reward Forcing: Efficient Streaming Video Generation with Rewarded Distribution Matching Distillation","ref_index":75,"is_internal_anchor":true},{"citing_arxiv_id":"2603.00918","citing_title":"Improving Text-to-Image Generation with Intrinsic Self-Confidence Rewards","ref_index":72,"is_internal_anchor":true},{"citing_arxiv_id":"2603.08090","citing_title":"DSH-Bench: A Difficulty- and Scenario-Aware Benchmark with Hierarchical Subject Taxonomy for Subject-Driven Text-to-Image Generation","ref_index":64,"is_internal_anchor":true},{"citing_arxiv_id":"2603.13224","citing_title":"Visual-ERM: Reward Modeling for Visual Equivalence","ref_index":33,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08063","citing_title":"Flow-OPD: On-Policy Distillation for Flow Matching Models","ref_index":46,"is_internal_anchor":true},{"citing_arxiv_id":"2509.16117","citing_title":"DiffusionNFT: Online Diffusion Reinforcement with Forward Process","ref_index":22,"is_internal_anchor":true},{"citing_arxiv_id":"2604.04142","citing_title":"OP-GRPO: Efficient Off-Policy GRPO for Flow-Matching Models","ref_index":39,"is_internal_anchor":true},{"citing_arxiv_id":"2507.21802","citing_title":"MixGRPO: Unlocking Flow-based GRPO Efficiency with Mixed ODE-SDE","ref_index":41,"is_internal_anchor":true},{"citing_arxiv_id":"2605.05922","citing_title":"Think, then Score: Decoupled Reasoning and Scoring for Video Reward Modeling","ref_index":41,"is_internal_anchor":true},{"citing_arxiv_id":"2605.11723","citing_title":"CaC: Advancing Video Reward Models via Hierarchical Spatiotemporal Concentrating","ref_index":53,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12112","citing_title":"When Policy Entropy Constraint Fails: Preserving Diversity in Flow-based RLHF via Perceptual Entropy","ref_index":65,"is_internal_anchor":true},{"citing_arxiv_id":"2502.17419","citing_title":"From System 1 to System 2: A Survey of Reasoning Large Language Models","ref_index":178,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08063","citing_title":"Flow-OPD: On-Policy Distillation for Flow Matching Models","ref_index":46,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08354","citing_title":"Auto-Rubric as Reward: From Implicit Preferences to Explicit Multimodal Generative Criteria","ref_index":39,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08703","citing_title":"RewardHarness: Self-Evolving Agentic Post-Training","ref_index":27,"is_internal_anchor":true},{"citing_arxiv_id":"2605.09291","citing_title":"dFlowGRPO: Rate-Aware Policy Optimization for Discrete Flow Models","ref_index":162,"is_internal_anchor":true},{"citing_arxiv_id":"2604.25427","citing_title":"A Systematic Post-Train Framework for Video Generation","ref_index":48,"is_internal_anchor":true},{"citing_arxiv_id":"2604.23380","citing_title":"V-GRPO: Online Reinforcement Learning for Denoising Generative Models Is Easier than You Think","ref_index":41,"is_internal_anchor":true},{"citing_arxiv_id":"2505.05470","citing_title":"Flow-GRPO: Training Flow Matching Models via Online RL","ref_index":61,"is_internal_anchor":true},{"citing_arxiv_id":"2604.22868","citing_title":"Probing Visual Planning in Image Editing Models","ref_index":78,"is_internal_anchor":true},{"citing_arxiv_id":"2604.19544","citing_title":"DT2IT-MRM: Debiased Preference Construction and Iterative Training for Multimodal Reward Modeling","ref_index":53,"is_internal_anchor":true},{"citing_arxiv_id":"2604.11490","citing_title":"Anthropogenic Regional Adaptation in Multimodal Vision-Language Model","ref_index":62,"is_internal_anchor":true},{"citing_arxiv_id":"2605.08063","citing_title":"Flow-OPD: On-Policy Distillation for Flow Matching Models","ref_index":46,"is_internal_anchor":true},{"citing_arxiv_id":"2605.07872","citing_title":"Video Understanding Reward Modeling: A Robust Benchmark and Performant Reward Models","ref_index":35,"is_internal_anchor":true},{"citing_arxiv_id":"2604.15311","citing_title":"LeapAlign: Post-Training Flow Matching Models at Any Generation Step by Building Two-Step Trajectories","ref_index":48,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":2,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/AWCZRZQ4QCF2K4O2MUR2KYFJ4P","json":"https://pith.science/pith/AWCZRZQ4QCF2K4O2MUR2KYFJ4P.json","graph_json":"https://pith.science/api/pith-number/AWCZRZQ4QCF2K4O2MUR2KYFJ4P/graph.json","events_json":"https://pith.science/api/pith-number/AWCZRZQ4QCF2K4O2MUR2KYFJ4P/events.json","paper":"https://pith.science/paper/AWCZRZQ4"},"agent_actions":{"view_html":"https://pith.science/pith/AWCZRZQ4QCF2K4O2MUR2KYFJ4P","download_json":"https://pith.science/pith/AWCZRZQ4QCF2K4O2MUR2KYFJ4P.json","view_paper":"https://pith.science/paper/AWCZRZQ4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2503.05236&json=true","fetch_graph":"https://pith.science/api/pith-number/AWCZRZQ4QCF2K4O2MUR2KYFJ4P/graph.json","fetch_events":"https://pith.science/api/pith-number/AWCZRZQ4QCF2K4O2MUR2KYFJ4P/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/AWCZRZQ4QCF2K4O2MUR2KYFJ4P/action/timestamp_anchor","attest_storage":"https://pith.science/pith/AWCZRZQ4QCF2K4O2MUR2KYFJ4P/action/storage_attestation","attest_author":"https://pith.science/pith/AWCZRZQ4QCF2K4O2MUR2KYFJ4P/action/author_attestation","sign_citation":"https://pith.science/pith/AWCZRZQ4QCF2K4O2MUR2KYFJ4P/action/citation_signature","submit_replication":"https://pith.science/pith/AWCZRZQ4QCF2K4O2MUR2KYFJ4P/action/replication_record"}},"created_at":"2026-05-18T03:15:18.210440+00:00","updated_at":"2026-05-18T03:15:18.210440+00:00"}