{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:2LABCZ5JCLRU5I5WG6PHH6VIRV","short_pith_number":"pith:2LABCZ5J","schema_version":"1.0","canonical_sha256":"d2c01167a912e34ea3b6379e73faa88d5169db81c8f3d7f0418490d4a5379e8f","source":{"kind":"arxiv","id":"2407.01392","version":4},"attestation_state":"computed","paper":{"title":"Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","cs.RO"],"primary_cat":"cs.LG","authors_text":"Boyuan Chen, Diego Marti Monso, Max Simchowitz, Russ Tedrake, Vincent Sitzmann, Yilun Du","submitted_at":"2024-07-01T15:43:25Z","abstract_excerpt":"This paper presents Diffusion Forcing, a new training paradigm where a diffusion model is trained to denoise a set of tokens with independent per-token noise levels. We apply Diffusion Forcing to sequence generative modeling by training a causal next-token prediction model to generate one or several future tokens without fully diffusing past ones. Our approach is shown to combine the strengths of next-token prediction models, such as variable-length generation, with the strengths of full-sequence diffusion models, such as the ability to guide sampling to desirable trajectories. Our method offe"},"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":"2407.01392","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-07-01T15:43:25Z","cross_cats_sorted":["cs.CV","cs.RO"],"title_canon_sha256":"b85ade2981091344596ce8e8580434c30596a6f5258f77c3bdbfa225bc971503","abstract_canon_sha256":"12df1f8740597c6a2aa9bf22b677911f40ebbca832f8f85c5ec1df8cc7e790e8"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:46:48.752530Z","signature_b64":"FLTd4Q5rFam8rham7LMvSzGzprpDxXo4xGUfYde43nSislJ8tR5EVKIaFxOu3CjRcs+/JYl1zSCP2bQqFptQDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"d2c01167a912e34ea3b6379e73faa88d5169db81c8f3d7f0418490d4a5379e8f","last_reissued_at":"2026-07-05T09:46:48.751996Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:46:48.751996Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Diffusion Forcing: Next-token Prediction Meets Full-Sequence Diffusion","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.CV","cs.RO"],"primary_cat":"cs.LG","authors_text":"Boyuan Chen, Diego Marti Monso, Max Simchowitz, Russ Tedrake, Vincent Sitzmann, Yilun Du","submitted_at":"2024-07-01T15:43:25Z","abstract_excerpt":"This paper presents Diffusion Forcing, a new training paradigm where a diffusion model is trained to denoise a set of tokens with independent per-token noise levels. We apply Diffusion Forcing to sequence generative modeling by training a causal next-token prediction model to generate one or several future tokens without fully diffusing past ones. Our approach is shown to combine the strengths of next-token prediction models, such as variable-length generation, with the strengths of full-sequence diffusion models, such as the ability to guide sampling to desirable trajectories. Our method offe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2407.01392","kind":"arxiv","version":4},"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/2407.01392/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":"2407.01392","created_at":"2026-07-05T09:46:48.752054+00:00"},{"alias_kind":"arxiv_version","alias_value":"2407.01392v4","created_at":"2026-07-05T09:46:48.752054+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2407.01392","created_at":"2026-07-05T09:46:48.752054+00:00"},{"alias_kind":"pith_short_12","alias_value":"2LABCZ5JCLRU","created_at":"2026-07-05T09:46:48.752054+00:00"},{"alias_kind":"pith_short_16","alias_value":"2LABCZ5JCLRU5I5W","created_at":"2026-07-05T09:46:48.752054+00:00"},{"alias_kind":"pith_short_8","alias_value":"2LABCZ5J","created_at":"2026-07-05T09:46:48.752054+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":20,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2607.08766","citing_title":"OPSD-V: On-Policy Self-Distillation for Post-Training Few-Step Autoregressive Video Generators","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2607.05376","citing_title":"MV-Forcing: Long Multi-View Video Generation via 4D-Grounded Spatio-Temporal Self-Forcing","ref_index":4,"is_internal_anchor":true},{"citing_arxiv_id":"2606.27364","citing_title":"PhysiFormer: Learning to Simulate Mechanics in World Space","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07079","citing_title":"AsyncPatch Diffusion: spatially-flexible image generation","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2606.31734","citing_title":"MemLearner: Learning to Query Context memory for Video World Models","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08729","citing_title":"Unison: Harmonizing Motion, Speech, and Sound for Human-Centric Audio-Video Generation","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.25659","citing_title":"StreamChar: Long-Horizon Streaming Character Audio-Video Generation with Decoupled Orchestration","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2606.00133","citing_title":"World Models: A Comprehensive Survey of Architectures, Methodologies, Reasoning Paradigms, and Applications","ref_index":294,"is_internal_anchor":false},{"citing_arxiv_id":"2602.16169","citing_title":"Discrete Stochastic Localization for Non-autoregressive Generation","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20299","citing_title":"Mechanisms of Misgeneralization in Physical Sequence Modeling","ref_index":104,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16520","citing_title":"Global Convergence of Sampling-Based Nonconvex Optimization through Diffusion-Style Smoothing","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2605.17019","citing_title":"StreamingEffect: Real-Time Human-Centric Video Effect Generation","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2502.10248","citing_title":"Step-Video-T2V Technical Report: The Practice, Challenges, and Future of Video Foundation Model","ref_index":54,"is_internal_anchor":false},{"citing_arxiv_id":"2509.18611","citing_title":"Flow marching for a generative PDE foundation model","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2408.14837","citing_title":"Diffusion Models Are Real-Time Game Engines","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2409.00588","citing_title":"Diffusion Policy Policy Optimization","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2602.11229","citing_title":"Latent Generative Solvers for Generalizable Long-Term Physics Simulation","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2603.10093","citing_title":"Equivariant Asynchronous Diffusion: An Adaptive Denoising Schedule for Accelerated Molecular Conformation Generation","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2504.02792","citing_title":"Unified World Models: Coupling Video and Action Diffusion for Pretraining on Large Robotic Datasets","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.08729","citing_title":"Unison: Harmonizing Motion, Speech, and Sound for Human-Centric Audio-Video Generation","ref_index":3,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/2LABCZ5JCLRU5I5WG6PHH6VIRV","json":"https://pith.science/pith/2LABCZ5JCLRU5I5WG6PHH6VIRV.json","graph_json":"https://pith.science/api/pith-number/2LABCZ5JCLRU5I5WG6PHH6VIRV/graph.json","events_json":"https://pith.science/api/pith-number/2LABCZ5JCLRU5I5WG6PHH6VIRV/events.json","paper":"https://pith.science/paper/2LABCZ5J"},"agent_actions":{"view_html":"https://pith.science/pith/2LABCZ5JCLRU5I5WG6PHH6VIRV","download_json":"https://pith.science/pith/2LABCZ5JCLRU5I5WG6PHH6VIRV.json","view_paper":"https://pith.science/paper/2LABCZ5J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2407.01392&json=true","fetch_graph":"https://pith.science/api/pith-number/2LABCZ5JCLRU5I5WG6PHH6VIRV/graph.json","fetch_events":"https://pith.science/api/pith-number/2LABCZ5JCLRU5I5WG6PHH6VIRV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/2LABCZ5JCLRU5I5WG6PHH6VIRV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/2LABCZ5JCLRU5I5WG6PHH6VIRV/action/storage_attestation","attest_author":"https://pith.science/pith/2LABCZ5JCLRU5I5WG6PHH6VIRV/action/author_attestation","sign_citation":"https://pith.science/pith/2LABCZ5JCLRU5I5WG6PHH6VIRV/action/citation_signature","submit_replication":"https://pith.science/pith/2LABCZ5JCLRU5I5WG6PHH6VIRV/action/replication_record"}},"created_at":"2026-07-05T09:46:48.752054+00:00","updated_at":"2026-07-05T09:46:48.752054+00:00"}