{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:OACBCVNT2A7QW7AB3OFT7VP4GK","short_pith_number":"pith:OACBCVNT","schema_version":"1.0","canonical_sha256":"70041155b3d03f0b7c01db8b3fd5fc32bcf338d6d9cafd5d61545d0836d3f134","source":{"kind":"arxiv","id":"2404.11599","version":1},"attestation_state":"computed","paper":{"title":"Variational Bayesian Last Layers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"James Harrison, Jasper Snoek, John Willes","submitted_at":"2024-04-17T17:50:24Z","abstract_excerpt":"We introduce a deterministic variational formulation for training Bayesian last layer neural networks. This yields a sampling-free, single-pass model and loss that effectively improves uncertainty estimation. Our variational Bayesian last layer (VBLL) can be trained and evaluated with only quadratic complexity in last layer width, and is thus (nearly) computationally free to add to standard architectures. We experimentally investigate VBLLs, and show that they improve predictive accuracy, calibration, and out of distribution detection over baselines across both regression and classification. F"},"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":"2404.11599","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-04-17T17:50:24Z","cross_cats_sorted":["cs.CV","stat.ML"],"title_canon_sha256":"9acf0894e41d1976376a7719b8b8ef4af6f6aa808e5561a5870b7783f574d98b","abstract_canon_sha256":"38fd3fd99122d7928c6d8456f0c9bcfe698e079d5849050082eaa3eefd1cfe98"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:09:10.693894Z","signature_b64":"D83Z0E8mJWgQ78tEaL5ngau+INEVEMrhUDCxLEkRJemhDBewYaGGvpjcjgLKdUemefPvbpBq/8K0M39s8O96DA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"70041155b3d03f0b7c01db8b3fd5fc32bcf338d6d9cafd5d61545d0836d3f134","last_reissued_at":"2026-07-05T08:09:10.693522Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:09:10.693522Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Variational Bayesian Last Layers","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CV","stat.ML"],"primary_cat":"cs.LG","authors_text":"James Harrison, Jasper Snoek, John Willes","submitted_at":"2024-04-17T17:50:24Z","abstract_excerpt":"We introduce a deterministic variational formulation for training Bayesian last layer neural networks. This yields a sampling-free, single-pass model and loss that effectively improves uncertainty estimation. Our variational Bayesian last layer (VBLL) can be trained and evaluated with only quadratic complexity in last layer width, and is thus (nearly) computationally free to add to standard architectures. We experimentally investigate VBLLs, and show that they improve predictive accuracy, calibration, and out of distribution detection over baselines across both regression and classification. F"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2404.11599","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/2404.11599/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":"2404.11599","created_at":"2026-07-05T08:09:10.693577+00:00"},{"alias_kind":"arxiv_version","alias_value":"2404.11599v1","created_at":"2026-07-05T08:09:10.693577+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2404.11599","created_at":"2026-07-05T08:09:10.693577+00:00"},{"alias_kind":"pith_short_12","alias_value":"OACBCVNT2A7Q","created_at":"2026-07-05T08:09:10.693577+00:00"},{"alias_kind":"pith_short_16","alias_value":"OACBCVNT2A7QW7AB","created_at":"2026-07-05T08:09:10.693577+00:00"},{"alias_kind":"pith_short_8","alias_value":"OACBCVNT","created_at":"2026-07-05T08:09:10.693577+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":7,"internal_anchor_count":2,"sample":[{"citing_arxiv_id":"2607.06776","citing_title":"Efficient Bayesian Deep Ensembles via Analytic Predictive Inference","ref_index":3,"is_internal_anchor":true},{"citing_arxiv_id":"2607.06943","citing_title":"General Incomplete Multimodal Learning via Dynamic Quality Perception","ref_index":13,"is_internal_anchor":true},{"citing_arxiv_id":"2606.07439","citing_title":"A 65 nm Multi-Modal Bayesian Inference Engine with 16.3 fJ/Sample Calibration-Free GRNG for Risk-Aware At-Home Skin Lesion Screening","ref_index":41,"is_internal_anchor":false},{"citing_arxiv_id":"2606.01595","citing_title":"Uncertainty-Calibrated Diffusion for Reliable 3D Molecular Graph Generation","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28203","citing_title":"Refining Multidimensional Video Reward Models via Disentangled Influence Functions","ref_index":32,"is_internal_anchor":false},{"citing_arxiv_id":"2605.14041","citing_title":"Wahkon: A Statistically Principled Deep RKHS Superposition Network","ref_index":36,"is_internal_anchor":false},{"citing_arxiv_id":"2605.12208","citing_title":"Self-Supervised Laplace Approximation for Bayesian Uncertainty Quantification","ref_index":44,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/OACBCVNT2A7QW7AB3OFT7VP4GK","json":"https://pith.science/pith/OACBCVNT2A7QW7AB3OFT7VP4GK.json","graph_json":"https://pith.science/api/pith-number/OACBCVNT2A7QW7AB3OFT7VP4GK/graph.json","events_json":"https://pith.science/api/pith-number/OACBCVNT2A7QW7AB3OFT7VP4GK/events.json","paper":"https://pith.science/paper/OACBCVNT"},"agent_actions":{"view_html":"https://pith.science/pith/OACBCVNT2A7QW7AB3OFT7VP4GK","download_json":"https://pith.science/pith/OACBCVNT2A7QW7AB3OFT7VP4GK.json","view_paper":"https://pith.science/paper/OACBCVNT","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2404.11599&json=true","fetch_graph":"https://pith.science/api/pith-number/OACBCVNT2A7QW7AB3OFT7VP4GK/graph.json","fetch_events":"https://pith.science/api/pith-number/OACBCVNT2A7QW7AB3OFT7VP4GK/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/OACBCVNT2A7QW7AB3OFT7VP4GK/action/timestamp_anchor","attest_storage":"https://pith.science/pith/OACBCVNT2A7QW7AB3OFT7VP4GK/action/storage_attestation","attest_author":"https://pith.science/pith/OACBCVNT2A7QW7AB3OFT7VP4GK/action/author_attestation","sign_citation":"https://pith.science/pith/OACBCVNT2A7QW7AB3OFT7VP4GK/action/citation_signature","submit_replication":"https://pith.science/pith/OACBCVNT2A7QW7AB3OFT7VP4GK/action/replication_record"}},"created_at":"2026-07-05T08:09:10.693577+00:00","updated_at":"2026-07-05T08:09:10.693577+00:00"}