{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:THBHO7MOYFM7L2BGAJAFRYU4JK","short_pith_number":"pith:THBHO7MO","schema_version":"1.0","canonical_sha256":"99c2777d8ec159f5e826024058e29c4a9ff709b3c6ee830210750eb8509cf17d","source":{"kind":"arxiv","id":"2605.21489","version":1},"attestation_state":"computed","paper":{"title":"Variance Reduction for Expectations with Diffusion Teachers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","stat.CO","stat.ML"],"primary_cat":"cs.LG","authors_text":"James Lucas, Jesse Bettencourt, Jonathan Lorraine, Matan Atzmon, Xindi Wu","submitted_at":"2026-05-20T17:59:52Z","abstract_excerpt":"Pretrained diffusion models serve as frozen teachers feeding downstream pipelines such as text-to-3D, single-step distillation, and data attribution. The teacher gradients these pipelines consume are Monte Carlo (MC) expectations over noise levels and Gaussian noise samples; their estimator variance dominates compute cost because each draw requires expensive upstream work (rendering, simulation, encoding). We introduce CARV, a compute-aware variance-accounting framework that motivates a hierarchical MC estimator: amortize the expensive upstream computation over cheap diffusion-noise resamples,"},"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":"2605.21489","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-20T17:59:52Z","cross_cats_sorted":["cs.AI","cs.CV","stat.CO","stat.ML"],"title_canon_sha256":"73a2d5177f96602bf72fb5fbd5fae8aa668bc418046b8fc2369e87a86198203d","abstract_canon_sha256":"2204816fa4edc3514aacf22f63033f48c1c919b2d68eff86d3289af6ed069a91"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-21T02:05:39.716766Z","signature_b64":"0GqfrWmfz2wWqohZpP9IdI/bR5JEsWiVC2E4RnuGhYJuS1hzO3KAn7BJLPDcue+0XrHPtwn3WeXAOA99XgdSDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"99c2777d8ec159f5e826024058e29c4a9ff709b3c6ee830210750eb8509cf17d","last_reissued_at":"2026-05-21T02:05:39.715955Z","signature_status":"signed_v1","first_computed_at":"2026-05-21T02:05:39.715955Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Variance Reduction for Expectations with Diffusion Teachers","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.CV","stat.CO","stat.ML"],"primary_cat":"cs.LG","authors_text":"James Lucas, Jesse Bettencourt, Jonathan Lorraine, Matan Atzmon, Xindi Wu","submitted_at":"2026-05-20T17:59:52Z","abstract_excerpt":"Pretrained diffusion models serve as frozen teachers feeding downstream pipelines such as text-to-3D, single-step distillation, and data attribution. The teacher gradients these pipelines consume are Monte Carlo (MC) expectations over noise levels and Gaussian noise samples; their estimator variance dominates compute cost because each draw requires expensive upstream work (rendering, simulation, encoding). We introduce CARV, a compute-aware variance-accounting framework that motivates a hierarchical MC estimator: amortize the expensive upstream computation over cheap diffusion-noise resamples,"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.21489","kind":"arxiv","version":1},"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/2605.21489/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":"2605.21489","created_at":"2026-05-21T02:05:39.716097+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.21489v1","created_at":"2026-05-21T02:05:39.716097+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.21489","created_at":"2026-05-21T02:05:39.716097+00:00"},{"alias_kind":"pith_short_12","alias_value":"THBHO7MOYFM7","created_at":"2026-05-21T02:05:39.716097+00:00"},{"alias_kind":"pith_short_16","alias_value":"THBHO7MOYFM7L2BG","created_at":"2026-05-21T02:05:39.716097+00:00"},{"alias_kind":"pith_short_8","alias_value":"THBHO7MO","created_at":"2026-05-21T02:05:39.716097+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/THBHO7MOYFM7L2BGAJAFRYU4JK","json":"https://pith.science/pith/THBHO7MOYFM7L2BGAJAFRYU4JK.json","graph_json":"https://pith.science/api/pith-number/THBHO7MOYFM7L2BGAJAFRYU4JK/graph.json","events_json":"https://pith.science/api/pith-number/THBHO7MOYFM7L2BGAJAFRYU4JK/events.json","paper":"https://pith.science/paper/THBHO7MO"},"agent_actions":{"view_html":"https://pith.science/pith/THBHO7MOYFM7L2BGAJAFRYU4JK","download_json":"https://pith.science/pith/THBHO7MOYFM7L2BGAJAFRYU4JK.json","view_paper":"https://pith.science/paper/THBHO7MO","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.21489&json=true","fetch_graph":"https://pith.science/api/pith-number/THBHO7MOYFM7L2BGAJAFRYU4JK/graph.json","fetch_events":"https://pith.science/api/pith-number/THBHO7MOYFM7L2BGAJAFRYU4JK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/THBHO7MOYFM7L2BGAJAFRYU4JK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/THBHO7MOYFM7L2BGAJAFRYU4JK/action/storage_attestation","attest_author":"https://pith.science/pith/THBHO7MOYFM7L2BGAJAFRYU4JK/action/author_attestation","sign_citation":"https://pith.science/pith/THBHO7MOYFM7L2BGAJAFRYU4JK/action/citation_signature","submit_replication":"https://pith.science/pith/THBHO7MOYFM7L2BGAJAFRYU4JK/action/replication_record"}},"created_at":"2026-05-21T02:05:39.716097+00:00","updated_at":"2026-05-21T02:05:39.716097+00:00"}