{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2020:RUNDNIGID5GUEU4IUZCVEPJIVH","short_pith_number":"pith:RUNDNIGI","schema_version":"1.0","canonical_sha256":"8d1a36a0c81f4d425388a645523d28a9def9ac4ad5f3aa0c4665d8c14cc426ff","source":{"kind":"arxiv","id":"2006.09011","version":2},"attestation_state":"computed","paper":{"title":"Improved Techniques for Training Score-Based Generative Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Stefano Ermon, Yang Song","submitted_at":"2020-06-16T09:17:17Z","abstract_excerpt":"Score-based generative models can produce high quality image samples comparable to GANs, without requiring adversarial optimization. However, existing training procedures are limited to images of low resolution (typically below 32x32), and can be unstable under some settings. We provide a new theoretical analysis of learning and sampling from score models in high dimensional spaces, explaining existing failure modes and motivating new solutions that generalize across datasets. To enhance stability, we also propose to maintain an exponential moving average of model weights. With these improveme"},"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":"2006.09011","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-06-16T09:17:17Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"55fbd73b291bcd30e73289ff03bbf31c248bd2f3b1f6c89eead7f1a5964e5fcb","abstract_canon_sha256":"d4f5c2cb8149b3bad9baacbc02888af881d953b43abb34d23a373fa509814528"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T01:45:43.380710Z","signature_b64":"vAG16sIQjVKDDv6kSnL5eWQs8vnsQq9MYD1KT6Ttp2FtJU+dtSm9afzhCbALMEeXoUJUE/E5kHvFlwNtwPHXCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8d1a36a0c81f4d425388a645523d28a9def9ac4ad5f3aa0c4665d8c14cc426ff","last_reissued_at":"2026-07-05T01:45:43.380256Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T01:45:43.380256Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Improved Techniques for Training Score-Based Generative Models","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"Stefano Ermon, Yang Song","submitted_at":"2020-06-16T09:17:17Z","abstract_excerpt":"Score-based generative models can produce high quality image samples comparable to GANs, without requiring adversarial optimization. However, existing training procedures are limited to images of low resolution (typically below 32x32), and can be unstable under some settings. We provide a new theoretical analysis of learning and sampling from score models in high dimensional spaces, explaining existing failure modes and motivating new solutions that generalize across datasets. To enhance stability, we also propose to maintain an exponential moving average of model weights. With these improveme"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2006.09011","kind":"arxiv","version":2},"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/2006.09011/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":"2006.09011","created_at":"2026-07-05T01:45:43.380312+00:00"},{"alias_kind":"arxiv_version","alias_value":"2006.09011v2","created_at":"2026-07-05T01:45:43.380312+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2006.09011","created_at":"2026-07-05T01:45:43.380312+00:00"},{"alias_kind":"pith_short_12","alias_value":"RUNDNIGID5GU","created_at":"2026-07-05T01:45:43.380312+00:00"},{"alias_kind":"pith_short_16","alias_value":"RUNDNIGID5GUEU4I","created_at":"2026-07-05T01:45:43.380312+00:00"},{"alias_kind":"pith_short_8","alias_value":"RUNDNIGI","created_at":"2026-07-05T01:45:43.380312+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":13,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.08505","citing_title":"Diffusion Models for Sampling Near Criticality in Lattice Field Theories","ref_index":59,"is_internal_anchor":true},{"citing_arxiv_id":"2605.27499","citing_title":"GenSBI: Generative Methods for Simulation-Based Inference in JAX","ref_index":59,"is_internal_anchor":false},{"citing_arxiv_id":"2010.02502","citing_title":"Denoising Diffusion Implicit Models","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2503.21432","citing_title":"Exploring the flavor structure of leptons via diffusion models","ref_index":40,"is_internal_anchor":false},{"citing_arxiv_id":"2101.02388","citing_title":"Knowledge Distillation in Iterative Generative Models for Improved Sampling Speed","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2102.09672","citing_title":"Improved Denoising Diffusion Probabilistic Models","ref_index":8,"is_internal_anchor":false},{"citing_arxiv_id":"2009.09761","citing_title":"DiffWave: A Versatile Diffusion Model for Audio Synthesis","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2105.05233","citing_title":"Diffusion Models Beat GANs on Image Synthesis","ref_index":58,"is_internal_anchor":false},{"citing_arxiv_id":"2108.01073","citing_title":"SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equations","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.06829","citing_title":"A Unified Measure-Theoretic View of Diffusion, Score-Based, and Flow Matching Generative Models","ref_index":13,"is_internal_anchor":false},{"citing_arxiv_id":"2006.11239","citing_title":"Denoising Diffusion Probabilistic Models","ref_index":56,"is_internal_anchor":false},{"citing_arxiv_id":"2311.15127","citing_title":"Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasets","ref_index":86,"is_internal_anchor":false},{"citing_arxiv_id":"2204.06125","citing_title":"Hierarchical Text-Conditional Image Generation with CLIP Latents","ref_index":49,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/RUNDNIGID5GUEU4IUZCVEPJIVH","json":"https://pith.science/pith/RUNDNIGID5GUEU4IUZCVEPJIVH.json","graph_json":"https://pith.science/api/pith-number/RUNDNIGID5GUEU4IUZCVEPJIVH/graph.json","events_json":"https://pith.science/api/pith-number/RUNDNIGID5GUEU4IUZCVEPJIVH/events.json","paper":"https://pith.science/paper/RUNDNIGI"},"agent_actions":{"view_html":"https://pith.science/pith/RUNDNIGID5GUEU4IUZCVEPJIVH","download_json":"https://pith.science/pith/RUNDNIGID5GUEU4IUZCVEPJIVH.json","view_paper":"https://pith.science/paper/RUNDNIGI","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2006.09011&json=true","fetch_graph":"https://pith.science/api/pith-number/RUNDNIGID5GUEU4IUZCVEPJIVH/graph.json","fetch_events":"https://pith.science/api/pith-number/RUNDNIGID5GUEU4IUZCVEPJIVH/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RUNDNIGID5GUEU4IUZCVEPJIVH/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RUNDNIGID5GUEU4IUZCVEPJIVH/action/storage_attestation","attest_author":"https://pith.science/pith/RUNDNIGID5GUEU4IUZCVEPJIVH/action/author_attestation","sign_citation":"https://pith.science/pith/RUNDNIGID5GUEU4IUZCVEPJIVH/action/citation_signature","submit_replication":"https://pith.science/pith/RUNDNIGID5GUEU4IUZCVEPJIVH/action/replication_record"}},"created_at":"2026-07-05T01:45:43.380312+00:00","updated_at":"2026-07-05T01:45:43.380312+00:00"}