{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:BONXDRYCSMUXFNDCYTM6TNI5AQ","short_pith_number":"pith:BONXDRYC","schema_version":"1.0","canonical_sha256":"0b9b71c702932972b462c4d9e9b51d0434b34cbc55657a9480213611358c7027","source":{"kind":"arxiv","id":"2207.13167","version":1},"attestation_state":"computed","paper":{"title":"One Simple Trick to Fix Your Bayesian Neural Network","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Ksawery Smoczy\\'nski, Marek Cygan, Philip Smolenski-Jensen, Piotr Tempczyk","submitted_at":"2022-07-26T19:45:36Z","abstract_excerpt":"One of the most popular estimation methods in Bayesian neural networks (BNN) is mean-field variational inference (MFVI). In this work, we show that neural networks with ReLU activation function induce posteriors, that are hard to fit with MFVI. We provide a theoretical justification for this phenomenon, study it empirically, and report the results of a series of experiments to investigate the effect of activation function on the calibration of BNNs. We find that using Leaky ReLU activations leads to more Gaussian-like weight posteriors and achieves a lower expected calibration error (ECE) than"},"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":"2207.13167","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"stat.ML","submitted_at":"2022-07-26T19:45:36Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"fe5cbaffece4d6259e3c2aba7a0e4f5fa2ee194707d0526db7205606893be2a7","abstract_canon_sha256":"b8251fa208aa8613d3ed9bfdc48aab6d5791fb35e2c2e4e601b85ee246caa98f"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:44:03.630801Z","signature_b64":"Xa71ONgKTPwL1Rww0BwyiGGJlOz9eVkZbesIJ0p1oajfCyjdNY2VUkwFhbjk/l2VOtZ4fw7LXL9YrNusyyLKDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0b9b71c702932972b462c4d9e9b51d0434b34cbc55657a9480213611358c7027","last_reissued_at":"2026-07-05T04:44:03.630410Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:44:03.630410Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"One Simple Trick to Fix Your Bayesian Neural Network","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Ksawery Smoczy\\'nski, Marek Cygan, Philip Smolenski-Jensen, Piotr Tempczyk","submitted_at":"2022-07-26T19:45:36Z","abstract_excerpt":"One of the most popular estimation methods in Bayesian neural networks (BNN) is mean-field variational inference (MFVI). In this work, we show that neural networks with ReLU activation function induce posteriors, that are hard to fit with MFVI. We provide a theoretical justification for this phenomenon, study it empirically, and report the results of a series of experiments to investigate the effect of activation function on the calibration of BNNs. We find that using Leaky ReLU activations leads to more Gaussian-like weight posteriors and achieves a lower expected calibration error (ECE) than"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2207.13167","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/2207.13167/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":"2207.13167","created_at":"2026-07-05T04:44:03.630466+00:00"},{"alias_kind":"arxiv_version","alias_value":"2207.13167v1","created_at":"2026-07-05T04:44:03.630466+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2207.13167","created_at":"2026-07-05T04:44:03.630466+00:00"},{"alias_kind":"pith_short_12","alias_value":"BONXDRYCSMUX","created_at":"2026-07-05T04:44:03.630466+00:00"},{"alias_kind":"pith_short_16","alias_value":"BONXDRYCSMUXFNDC","created_at":"2026-07-05T04:44:03.630466+00:00"},{"alias_kind":"pith_short_8","alias_value":"BONXDRYC","created_at":"2026-07-05T04:44:03.630466+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/BONXDRYCSMUXFNDCYTM6TNI5AQ","json":"https://pith.science/pith/BONXDRYCSMUXFNDCYTM6TNI5AQ.json","graph_json":"https://pith.science/api/pith-number/BONXDRYCSMUXFNDCYTM6TNI5AQ/graph.json","events_json":"https://pith.science/api/pith-number/BONXDRYCSMUXFNDCYTM6TNI5AQ/events.json","paper":"https://pith.science/paper/BONXDRYC"},"agent_actions":{"view_html":"https://pith.science/pith/BONXDRYCSMUXFNDCYTM6TNI5AQ","download_json":"https://pith.science/pith/BONXDRYCSMUXFNDCYTM6TNI5AQ.json","view_paper":"https://pith.science/paper/BONXDRYC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2207.13167&json=true","fetch_graph":"https://pith.science/api/pith-number/BONXDRYCSMUXFNDCYTM6TNI5AQ/graph.json","fetch_events":"https://pith.science/api/pith-number/BONXDRYCSMUXFNDCYTM6TNI5AQ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BONXDRYCSMUXFNDCYTM6TNI5AQ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BONXDRYCSMUXFNDCYTM6TNI5AQ/action/storage_attestation","attest_author":"https://pith.science/pith/BONXDRYCSMUXFNDCYTM6TNI5AQ/action/author_attestation","sign_citation":"https://pith.science/pith/BONXDRYCSMUXFNDCYTM6TNI5AQ/action/citation_signature","submit_replication":"https://pith.science/pith/BONXDRYCSMUXFNDCYTM6TNI5AQ/action/replication_record"}},"created_at":"2026-07-05T04:44:03.630466+00:00","updated_at":"2026-07-05T04:44:03.630466+00:00"}