{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:TGNYGJIQPWITOFFOZX3GAMXSJM","short_pith_number":"pith:TGNYGJIQ","schema_version":"1.0","canonical_sha256":"999b8325107d913714aecdf66032f24b1590db16b1f0fcd99b7c9fc0fef7123b","source":{"kind":"arxiv","id":"2102.01691","version":1},"attestation_state":"computed","paper":{"title":"Exact Langevin Dynamics with Stochastic Gradients","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Adri\\`a Garriga-Alonso, Vincent Fortuin","submitted_at":"2021-02-02T18:59:31Z","abstract_excerpt":"Stochastic gradient Markov Chain Monte Carlo algorithms are popular samplers for approximate inference, but they are generally biased. We show that many recent versions of these methods (e.g. Chen et al. (2014)) cannot be corrected using Metropolis-Hastings rejection sampling, because their acceptance probability is always zero. We can fix this by employing a sampler with realizable backwards trajectories, such as Gradient-Guided Monte Carlo (Horowitz, 1991), which generalizes stochastic gradient Langevin dynamics (Welling and Teh, 2011) and Hamiltonian Monte Carlo. We show that this sampler c"},"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":"2102.01691","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-sa/4.0/","primary_cat":"stat.ML","submitted_at":"2021-02-02T18:59:31Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"01800179ce15aa5409e916cea36b300b8e57c7c83c53e482a6cc4f7813b10e4d","abstract_canon_sha256":"190d7cab2c494e0972aef6cb0190f731f88c7978998b4938b4cf305e4689454c"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:11:32.694531Z","signature_b64":"FmEhMPWo9t+6q9Ns5wbRueyMQY0jTUtnekbcgwL6weJcgz7LWvNJybp40VdMAAh6+28lko1vTWtbJOL8K8xYAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"999b8325107d913714aecdf66032f24b1590db16b1f0fcd99b7c9fc0fef7123b","last_reissued_at":"2026-07-05T02:11:32.694117Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:11:32.694117Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Exact Langevin Dynamics with Stochastic Gradients","license":"http://creativecommons.org/licenses/by-sa/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Adri\\`a Garriga-Alonso, Vincent Fortuin","submitted_at":"2021-02-02T18:59:31Z","abstract_excerpt":"Stochastic gradient Markov Chain Monte Carlo algorithms are popular samplers for approximate inference, but they are generally biased. We show that many recent versions of these methods (e.g. Chen et al. (2014)) cannot be corrected using Metropolis-Hastings rejection sampling, because their acceptance probability is always zero. We can fix this by employing a sampler with realizable backwards trajectories, such as Gradient-Guided Monte Carlo (Horowitz, 1991), which generalizes stochastic gradient Langevin dynamics (Welling and Teh, 2011) and Hamiltonian Monte Carlo. We show that this sampler c"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.01691","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/2102.01691/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":"2102.01691","created_at":"2026-07-05T02:11:32.694181+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.01691v1","created_at":"2026-07-05T02:11:32.694181+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.01691","created_at":"2026-07-05T02:11:32.694181+00:00"},{"alias_kind":"pith_short_12","alias_value":"TGNYGJIQPWIT","created_at":"2026-07-05T02:11:32.694181+00:00"},{"alias_kind":"pith_short_16","alias_value":"TGNYGJIQPWITOFFO","created_at":"2026-07-05T02:11:32.694181+00:00"},{"alias_kind":"pith_short_8","alias_value":"TGNYGJIQ","created_at":"2026-07-05T02:11:32.694181+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.21488","citing_title":"Priors Matter: Addressing Misspecification in Bayesian Deep Q-Learning","ref_index":16,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/TGNYGJIQPWITOFFOZX3GAMXSJM","json":"https://pith.science/pith/TGNYGJIQPWITOFFOZX3GAMXSJM.json","graph_json":"https://pith.science/api/pith-number/TGNYGJIQPWITOFFOZX3GAMXSJM/graph.json","events_json":"https://pith.science/api/pith-number/TGNYGJIQPWITOFFOZX3GAMXSJM/events.json","paper":"https://pith.science/paper/TGNYGJIQ"},"agent_actions":{"view_html":"https://pith.science/pith/TGNYGJIQPWITOFFOZX3GAMXSJM","download_json":"https://pith.science/pith/TGNYGJIQPWITOFFOZX3GAMXSJM.json","view_paper":"https://pith.science/paper/TGNYGJIQ","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.01691&json=true","fetch_graph":"https://pith.science/api/pith-number/TGNYGJIQPWITOFFOZX3GAMXSJM/graph.json","fetch_events":"https://pith.science/api/pith-number/TGNYGJIQPWITOFFOZX3GAMXSJM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TGNYGJIQPWITOFFOZX3GAMXSJM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TGNYGJIQPWITOFFOZX3GAMXSJM/action/storage_attestation","attest_author":"https://pith.science/pith/TGNYGJIQPWITOFFOZX3GAMXSJM/action/author_attestation","sign_citation":"https://pith.science/pith/TGNYGJIQPWITOFFOZX3GAMXSJM/action/citation_signature","submit_replication":"https://pith.science/pith/TGNYGJIQPWITOFFOZX3GAMXSJM/action/replication_record"}},"created_at":"2026-07-05T02:11:32.694181+00:00","updated_at":"2026-07-05T02:11:32.694181+00:00"}