{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2023:QFDWOKK2WH3K4JJYIO56YV5X7E","short_pith_number":"pith:QFDWOKK2","schema_version":"1.0","canonical_sha256":"814767295ab1f6ae253843bbec57b7f937fc7ebafe44f5d0d002512b64000f93","source":{"kind":"arxiv","id":"2305.15583","version":8},"attestation_state":"computed","paper":{"title":"Alleviating Exposure Bias in Diffusion Models through Sampling with Shifted Time Steps","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Marie-Francine Moens, Mingxiao Li, Ruicong Yao, Tingyu Qu, Wei Sun","submitted_at":"2023-05-24T21:39:27Z","abstract_excerpt":"Diffusion Probabilistic Models (DPM) have shown remarkable efficacy in the synthesis of high-quality images. However, their inference process characteristically requires numerous, potentially hundreds, of iterative steps, which could exaggerate the problem of exposure bias due to the training and inference discrepancy. Previous work has attempted to mitigate this issue by perturbing inputs during training, which consequently mandates the retraining of the DPM. In this work, we conduct a systematic study of exposure bias in DPM and, intriguingly, we find that the exposure bias could be alleviat"},"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":"2305.15583","kind":"arxiv","version":8},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.CV","submitted_at":"2023-05-24T21:39:27Z","cross_cats_sorted":[],"title_canon_sha256":"a78e4d1410f2397210e3c65ac661fbbe003c1f861eb74a4268b05cac84615d29","abstract_canon_sha256":"ca8a732200ec4fa4966491936f81b9ea65c0bcd159bba4f34d278693826e947b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:32:29.779119Z","signature_b64":"vuPfjVmUEgg4ro3SOoeSjX0a2g0NTEORbttQpZ7IokSdXp0cFRvdN8iOIkqoh6G9dXBS5gFAqmOgU7gB146RCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"814767295ab1f6ae253843bbec57b7f937fc7ebafe44f5d0d002512b64000f93","last_reissued_at":"2026-07-05T08:32:29.778647Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:32:29.778647Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Alleviating Exposure Bias in Diffusion Models through Sampling with Shifted Time Steps","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Marie-Francine Moens, Mingxiao Li, Ruicong Yao, Tingyu Qu, Wei Sun","submitted_at":"2023-05-24T21:39:27Z","abstract_excerpt":"Diffusion Probabilistic Models (DPM) have shown remarkable efficacy in the synthesis of high-quality images. However, their inference process characteristically requires numerous, potentially hundreds, of iterative steps, which could exaggerate the problem of exposure bias due to the training and inference discrepancy. Previous work has attempted to mitigate this issue by perturbing inputs during training, which consequently mandates the retraining of the DPM. In this work, we conduct a systematic study of exposure bias in DPM and, intriguingly, we find that the exposure bias could be alleviat"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2305.15583","kind":"arxiv","version":8},"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/2305.15583/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":"2305.15583","created_at":"2026-07-05T08:32:29.778703+00:00"},{"alias_kind":"arxiv_version","alias_value":"2305.15583v8","created_at":"2026-07-05T08:32:29.778703+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2305.15583","created_at":"2026-07-05T08:32:29.778703+00:00"},{"alias_kind":"pith_short_12","alias_value":"QFDWOKK2WH3K","created_at":"2026-07-05T08:32:29.778703+00:00"},{"alias_kind":"pith_short_16","alias_value":"QFDWOKK2WH3K4JJY","created_at":"2026-07-05T08:32:29.778703+00:00"},{"alias_kind":"pith_short_8","alias_value":"QFDWOKK2","created_at":"2026-07-05T08:32:29.778703+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":3,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.19162","citing_title":"The Reward Was in Your Data All Along: Correcting Flow Matching with Discriminator-Guided RL","ref_index":124,"is_internal_anchor":false},{"citing_arxiv_id":"2510.02307","citing_title":"NoiseShift: Resolution-Aware Noise Recalibration for Better Low-Resolution Image Generation","ref_index":21,"is_internal_anchor":false},{"citing_arxiv_id":"2602.07775","citing_title":"Rolling Sink: Bridging Limited-Horizon Training and Open-Ended Testing in Autoregressive Video Diffusion","ref_index":57,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/QFDWOKK2WH3K4JJYIO56YV5X7E","json":"https://pith.science/pith/QFDWOKK2WH3K4JJYIO56YV5X7E.json","graph_json":"https://pith.science/api/pith-number/QFDWOKK2WH3K4JJYIO56YV5X7E/graph.json","events_json":"https://pith.science/api/pith-number/QFDWOKK2WH3K4JJYIO56YV5X7E/events.json","paper":"https://pith.science/paper/QFDWOKK2"},"agent_actions":{"view_html":"https://pith.science/pith/QFDWOKK2WH3K4JJYIO56YV5X7E","download_json":"https://pith.science/pith/QFDWOKK2WH3K4JJYIO56YV5X7E.json","view_paper":"https://pith.science/paper/QFDWOKK2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2305.15583&json=true","fetch_graph":"https://pith.science/api/pith-number/QFDWOKK2WH3K4JJYIO56YV5X7E/graph.json","fetch_events":"https://pith.science/api/pith-number/QFDWOKK2WH3K4JJYIO56YV5X7E/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/QFDWOKK2WH3K4JJYIO56YV5X7E/action/timestamp_anchor","attest_storage":"https://pith.science/pith/QFDWOKK2WH3K4JJYIO56YV5X7E/action/storage_attestation","attest_author":"https://pith.science/pith/QFDWOKK2WH3K4JJYIO56YV5X7E/action/author_attestation","sign_citation":"https://pith.science/pith/QFDWOKK2WH3K4JJYIO56YV5X7E/action/citation_signature","submit_replication":"https://pith.science/pith/QFDWOKK2WH3K4JJYIO56YV5X7E/action/replication_record"}},"created_at":"2026-07-05T08:32:29.778703+00:00","updated_at":"2026-07-05T08:32:29.778703+00:00"}