{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:WNBT3XSUDGYHN4BTLIU7TFTBQZ","short_pith_number":"pith:WNBT3XSU","schema_version":"1.0","canonical_sha256":"b3433dde5419b076f0335a29f996618678380ccb2fe394bdc677f50b003d7fb2","source":{"kind":"arxiv","id":"2402.14758","version":2},"attestation_state":"computed","paper":{"title":"Batch and match: black-box variational inference with a score-based divergence","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.CO"],"primary_cat":"stat.ML","authors_text":"Charles C. Margossian, Chirag Modi, David M. Blei, Diana Cai, Lawrence K. Saul, Loucas Pillaud-Vivien, Robert M. Gower","submitted_at":"2024-02-22T18:20:22Z","abstract_excerpt":"Most leading implementations of black-box variational inference (BBVI) are based on optimizing a stochastic evidence lower bound (ELBO). But such approaches to BBVI often converge slowly due to the high variance of their gradient estimates and their sensitivity to hyperparameters. In this work, we propose batch and match (BaM), an alternative approach to BBVI based on a score-based divergence. Notably, this score-based divergence can be optimized by a closed-form proximal update for Gaussian variational families with full covariance matrices. We analyze the convergence of BaM when the target d"},"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":"2402.14758","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2024-02-22T18:20:22Z","cross_cats_sorted":["cs.AI","cs.LG","stat.CO"],"title_canon_sha256":"401bf7dea5bbbbd19626448e66db8e4822d5badcef20a37c93a9798668917b84","abstract_canon_sha256":"f7ea3e36987a74e2a963ab84df774c54df05899d4a8cc645da546d7d0df571fe"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:30:50.511852Z","signature_b64":"GiDaSXIDjUoHTumBtp8wZ7gIdztEwmHkeNS14zf3fqcZMtXcpULcgjVEIVwE0NFuAXDJf8RTl7ou2KCFa0CwBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b3433dde5419b076f0335a29f996618678380ccb2fe394bdc677f50b003d7fb2","last_reissued_at":"2026-07-05T08:30:50.511364Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:30:50.511364Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Batch and match: black-box variational inference with a score-based divergence","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG","stat.CO"],"primary_cat":"stat.ML","authors_text":"Charles C. Margossian, Chirag Modi, David M. Blei, Diana Cai, Lawrence K. Saul, Loucas Pillaud-Vivien, Robert M. Gower","submitted_at":"2024-02-22T18:20:22Z","abstract_excerpt":"Most leading implementations of black-box variational inference (BBVI) are based on optimizing a stochastic evidence lower bound (ELBO). But such approaches to BBVI often converge slowly due to the high variance of their gradient estimates and their sensitivity to hyperparameters. In this work, we propose batch and match (BaM), an alternative approach to BBVI based on a score-based divergence. Notably, this score-based divergence can be optimized by a closed-form proximal update for Gaussian variational families with full covariance matrices. We analyze the convergence of BaM when the target d"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.14758","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/2402.14758/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":"2402.14758","created_at":"2026-07-05T08:30:50.511421+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.14758v2","created_at":"2026-07-05T08:30:50.511421+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.14758","created_at":"2026-07-05T08:30:50.511421+00:00"},{"alias_kind":"pith_short_12","alias_value":"WNBT3XSUDGYH","created_at":"2026-07-05T08:30:50.511421+00:00"},{"alias_kind":"pith_short_16","alias_value":"WNBT3XSUDGYHN4BT","created_at":"2026-07-05T08:30:50.511421+00:00"},{"alias_kind":"pith_short_8","alias_value":"WNBT3XSU","created_at":"2026-07-05T08:30:50.511421+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2605.07531","citing_title":"SGD for Variational Inference: Tackling Unbounded Variance via Preconditioning and Dynamic Batching","ref_index":9,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/WNBT3XSUDGYHN4BTLIU7TFTBQZ","json":"https://pith.science/pith/WNBT3XSUDGYHN4BTLIU7TFTBQZ.json","graph_json":"https://pith.science/api/pith-number/WNBT3XSUDGYHN4BTLIU7TFTBQZ/graph.json","events_json":"https://pith.science/api/pith-number/WNBT3XSUDGYHN4BTLIU7TFTBQZ/events.json","paper":"https://pith.science/paper/WNBT3XSU"},"agent_actions":{"view_html":"https://pith.science/pith/WNBT3XSUDGYHN4BTLIU7TFTBQZ","download_json":"https://pith.science/pith/WNBT3XSUDGYHN4BTLIU7TFTBQZ.json","view_paper":"https://pith.science/paper/WNBT3XSU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.14758&json=true","fetch_graph":"https://pith.science/api/pith-number/WNBT3XSUDGYHN4BTLIU7TFTBQZ/graph.json","fetch_events":"https://pith.science/api/pith-number/WNBT3XSUDGYHN4BTLIU7TFTBQZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WNBT3XSUDGYHN4BTLIU7TFTBQZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WNBT3XSUDGYHN4BTLIU7TFTBQZ/action/storage_attestation","attest_author":"https://pith.science/pith/WNBT3XSUDGYHN4BTLIU7TFTBQZ/action/author_attestation","sign_citation":"https://pith.science/pith/WNBT3XSUDGYHN4BTLIU7TFTBQZ/action/citation_signature","submit_replication":"https://pith.science/pith/WNBT3XSUDGYHN4BTLIU7TFTBQZ/action/replication_record"}},"created_at":"2026-07-05T08:30:50.511421+00:00","updated_at":"2026-07-05T08:30:50.511421+00:00"}