{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:PYG2Z6WYPLTNTS6KWHK73CPPL6","short_pith_number":"pith:PYG2Z6WY","schema_version":"1.0","canonical_sha256":"7e0dacfad87ae6d9cbcab1d5fd89ef5f990173206f469d4edce88f8d9d1e4afe","source":{"kind":"arxiv","id":"2301.13362","version":4},"attestation_state":"computed","paper":{"title":"Optimizing DDPM Sampling with Shortcut Fine-Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Kangwook Lee, Ying Fan","submitted_at":"2023-01-31T01:37:48Z","abstract_excerpt":"In this study, we propose Shortcut Fine-Tuning (SFT), a new approach for addressing the challenge of fast sampling of pretrained Denoising Diffusion Probabilistic Models (DDPMs). SFT advocates for the fine-tuning of DDPM samplers through the direct minimization of Integral Probability Metrics (IPM), instead of learning the backward diffusion process. This enables samplers to discover an alternative and more efficient sampling shortcut, deviating from the backward diffusion process. Inspired by a control perspective, we propose a new algorithm SFT-PG: Shortcut Fine-Tuning with Policy Gradient, "},"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":"2301.13362","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2023-01-31T01:37:48Z","cross_cats_sorted":[],"title_canon_sha256":"e9005cef04a88f2284a6e5148c22c8e7b63c4cfe0540127c8ee70facd4cd8785","abstract_canon_sha256":"fdb293454129fee83c3d6f6ea7d6c032d2db1563d3d72287aba758dfef6d2727"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:09:16.678840Z","signature_b64":"uyM0GF2A36PcTuU69sGFqZX71xFVLdBEq5HFCFB8Lu6V1+56haIiJhNpw8+jPHQq2FBVaF5OlhyudeeHIzCTDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"7e0dacfad87ae6d9cbcab1d5fd89ef5f990173206f469d4edce88f8d9d1e4afe","last_reissued_at":"2026-07-05T09:09:16.678308Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:09:16.678308Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Optimizing DDPM Sampling with Shortcut Fine-Tuning","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Kangwook Lee, Ying Fan","submitted_at":"2023-01-31T01:37:48Z","abstract_excerpt":"In this study, we propose Shortcut Fine-Tuning (SFT), a new approach for addressing the challenge of fast sampling of pretrained Denoising Diffusion Probabilistic Models (DDPMs). SFT advocates for the fine-tuning of DDPM samplers through the direct minimization of Integral Probability Metrics (IPM), instead of learning the backward diffusion process. This enables samplers to discover an alternative and more efficient sampling shortcut, deviating from the backward diffusion process. Inspired by a control perspective, we propose a new algorithm SFT-PG: Shortcut Fine-Tuning with Policy Gradient, "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2301.13362","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/2301.13362/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":"2301.13362","created_at":"2026-07-05T09:09:16.678370+00:00"},{"alias_kind":"arxiv_version","alias_value":"2301.13362v4","created_at":"2026-07-05T09:09:16.678370+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2301.13362","created_at":"2026-07-05T09:09:16.678370+00:00"},{"alias_kind":"pith_short_12","alias_value":"PYG2Z6WYPLTN","created_at":"2026-07-05T09:09:16.678370+00:00"},{"alias_kind":"pith_short_16","alias_value":"PYG2Z6WYPLTNTS6K","created_at":"2026-07-05T09:09:16.678370+00:00"},{"alias_kind":"pith_short_8","alias_value":"PYG2Z6WY","created_at":"2026-07-05T09:09:16.678370+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":13,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.22394","citing_title":"Curvature-Adaptive Consistency Flow Matching: Autonomous Trajectory Optimization via Reinforcement Learning","ref_index":9,"is_internal_anchor":false},{"citing_arxiv_id":"2607.02291","citing_title":"Optimizing Visual Generative Models via Distribution-wise Rewards","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2607.00486","citing_title":"PAPA: Online Personalized Active Preference Alignment","ref_index":7,"is_internal_anchor":false},{"citing_arxiv_id":"2511.18719","citing_title":"Seeing What Matters: Visual Preference Policy Optimization for Visual Generation","ref_index":3,"is_internal_anchor":false},{"citing_arxiv_id":"2605.20209","citing_title":"NaP-Control: Navigating Diffusion Prior for Versatile and Fast Character Control","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2605.15803","citing_title":"Embedding-perturbed Exploration Preference Optimization for Flow Models","ref_index":15,"is_internal_anchor":false},{"citing_arxiv_id":"2512.01236","citing_title":"PSR: Scaling Multi-Subject Personalized Image Generation with Pairwise Subject-Consistency Rewards","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2409.00588","citing_title":"Diffusion Policy Policy Optimization","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2507.21802","citing_title":"MixGRPO: Unlocking Flow-based GRPO Efficiency with Mixed ODE-SDE","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23380","citing_title":"V-GRPO: Online Reinforcement Learning for Denoising Generative Models Is Easier than You Think","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2305.13301","citing_title":"Training Diffusion Models with Reinforcement Learning","ref_index":5,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06583","citing_title":"Improved techniques for fine-tuning flow models via adjoint matching: a deterministic control pipeline","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2604.19234","citing_title":"Learning to Credit the Right Steps: Objective-aware Process Optimization for Visual Generation","ref_index":7,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/PYG2Z6WYPLTNTS6KWHK73CPPL6","json":"https://pith.science/pith/PYG2Z6WYPLTNTS6KWHK73CPPL6.json","graph_json":"https://pith.science/api/pith-number/PYG2Z6WYPLTNTS6KWHK73CPPL6/graph.json","events_json":"https://pith.science/api/pith-number/PYG2Z6WYPLTNTS6KWHK73CPPL6/events.json","paper":"https://pith.science/paper/PYG2Z6WY"},"agent_actions":{"view_html":"https://pith.science/pith/PYG2Z6WYPLTNTS6KWHK73CPPL6","download_json":"https://pith.science/pith/PYG2Z6WYPLTNTS6KWHK73CPPL6.json","view_paper":"https://pith.science/paper/PYG2Z6WY","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2301.13362&json=true","fetch_graph":"https://pith.science/api/pith-number/PYG2Z6WYPLTNTS6KWHK73CPPL6/graph.json","fetch_events":"https://pith.science/api/pith-number/PYG2Z6WYPLTNTS6KWHK73CPPL6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/PYG2Z6WYPLTNTS6KWHK73CPPL6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/PYG2Z6WYPLTNTS6KWHK73CPPL6/action/storage_attestation","attest_author":"https://pith.science/pith/PYG2Z6WYPLTNTS6KWHK73CPPL6/action/author_attestation","sign_citation":"https://pith.science/pith/PYG2Z6WYPLTNTS6KWHK73CPPL6/action/citation_signature","submit_replication":"https://pith.science/pith/PYG2Z6WYPLTNTS6KWHK73CPPL6/action/replication_record"}},"created_at":"2026-07-05T09:09:16.678370+00:00","updated_at":"2026-07-05T09:09:16.678370+00:00"}