{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:IAOFXIVIEKBQME3X7PBLCGOTTX","short_pith_number":"pith:IAOFXIVI","canonical_record":{"source":{"id":"2410.16901","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-22T11:15:07Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"0621ead406431bf076b3c359eed55de5f726c6f3b68fc4089ef0e44cb642f028","abstract_canon_sha256":"35b81f1318fa48e7b81ac9fd4d070ef71a77385c402f78c05c1897f66a100c96"},"schema_version":"1.0"},"canonical_sha256":"401c5ba2a82283061377fbc2b119d39dd3a1e19fcc3ae29ed26f2567cbb5b81b","source":{"kind":"arxiv","id":"2410.16901","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.16901","created_at":"2026-07-05T09:24:08Z"},{"alias_kind":"arxiv_version","alias_value":"2410.16901v1","created_at":"2026-07-05T09:24:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.16901","created_at":"2026-07-05T09:24:08Z"},{"alias_kind":"pith_short_12","alias_value":"IAOFXIVIEKBQ","created_at":"2026-07-05T09:24:08Z"},{"alias_kind":"pith_short_16","alias_value":"IAOFXIVIEKBQME3X","created_at":"2026-07-05T09:24:08Z"},{"alias_kind":"pith_short_8","alias_value":"IAOFXIVI","created_at":"2026-07-05T09:24:08Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:IAOFXIVIEKBQME3X7PBLCGOTTX","target":"record","payload":{"canonical_record":{"source":{"id":"2410.16901","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-22T11:15:07Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"0621ead406431bf076b3c359eed55de5f726c6f3b68fc4089ef0e44cb642f028","abstract_canon_sha256":"35b81f1318fa48e7b81ac9fd4d070ef71a77385c402f78c05c1897f66a100c96"},"schema_version":"1.0"},"canonical_sha256":"401c5ba2a82283061377fbc2b119d39dd3a1e19fcc3ae29ed26f2567cbb5b81b","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:24:08.382155Z","signature_b64":"YF39q56rLUG0eGzPKs/QlfBcsQl7JGZLvjriCSZnxA5t+lzjylAG2M6ZSokrcwpvuwTWDYoorg1m8KC5LGdsDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"401c5ba2a82283061377fbc2b119d39dd3a1e19fcc3ae29ed26f2567cbb5b81b","last_reissued_at":"2026-07-05T09:24:08.381659Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:24:08.381659Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2410.16901","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:24:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"xqsn0zPZoxxRSRHvKRNdkobRBge9AQCG/lgI/NpQ96Ei/3g4wcvcv8XIEGkZijNCFguW48TN0PQFLLnFjEE6Cg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-23T17:37:56.306841Z"},"content_sha256":"f5af3cefabe68b801bbe0d24683b7441f5f355795a3708a596e5f974f9d39510","schema_version":"1.0","event_id":"sha256:f5af3cefabe68b801bbe0d24683b7441f5f355795a3708a596e5f974f9d39510"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:IAOFXIVIEKBQME3X7PBLCGOTTX","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Bayes without Underfitting: Fully Correlated Deep Learning Posteriors via Alternating Projections","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Hrittik Roy, Marco Miani, S{\\o}ren Hauberg","submitted_at":"2024-10-22T11:15:07Z","abstract_excerpt":"Bayesian deep learning all too often underfits so that the Bayesian prediction is less accurate than a simple point estimate. Uncertainty quantification then comes at the cost of accuracy. For linearized models, the null space of the generalized Gauss-Newton matrix corresponds to parameters that preserve the training predictions of the point estimate. We propose to build Bayesian approximations in this null space, thereby guaranteeing that the Bayesian predictive does not underfit. We suggest a matrix-free algorithm for projecting onto this null space, which scales linearly with the number of "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.16901","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/2410.16901/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"},"verdict_id":null},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T09:24:08Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"0Lk50E7pu1IZPLsEQbMxE84GcztXu73yvGNDW/6BhiWIf2whpP0CzXS+2PM093wCquq6CdLZALbQghBBGzjjBg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-23T17:37:56.307214Z"},"content_sha256":"8c6beadc56bb22289ca0a5afe8d36d98a72ab0d9dc18ebe881f9e8541015c13c","schema_version":"1.0","event_id":"sha256:8c6beadc56bb22289ca0a5afe8d36d98a72ab0d9dc18ebe881f9e8541015c13c"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/IAOFXIVIEKBQME3X7PBLCGOTTX/bundle.json","state_url":"https://pith.science/pith/IAOFXIVIEKBQME3X7PBLCGOTTX/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/IAOFXIVIEKBQME3X7PBLCGOTTX/bundle.json","status":"primary"}],"public_keys":[{"key_id":"pith-v1-2026-05","algorithm":"ed25519","format":"raw","public_key_b64":"stVStoiQhXFxp4s2pdzPNoqVNBMojDU/fJ2db5S3CbM=","public_key_hex":"b2d552b68890857171a78b36a5dccf368a953413288c353f7c9d9d6f94b709b3","fingerprint_sha256_b32_first128bits":"RVFV5Z2OI2J3ZUO7ERDEBCYNKS","fingerprint_sha256_hex":"8d4b5ee74e4693bcd1df2446408b0d54","rotates_at":null,"url":"https://pith.science/pith-signing-key.json","notes":"Pith uses this Ed25519 key to sign canonical record SHA-256 digests. Verify with: ed25519_verify(public_key, message=canonical_sha256_bytes, signature=base64decode(signature_b64))."}],"merge_version":"pith-open-graph-merge-v1","built_at":"2026-08-23T17:37:56Z","links":{"resolver":"https://pith.science/pith/IAOFXIVIEKBQME3X7PBLCGOTTX","bundle":"https://pith.science/pith/IAOFXIVIEKBQME3X7PBLCGOTTX/bundle.json","state":"https://pith.science/pith/IAOFXIVIEKBQME3X7PBLCGOTTX/state.json","well_known_bundle":"https://pith.science/.well-known/pith/IAOFXIVIEKBQME3X7PBLCGOTTX/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:IAOFXIVIEKBQME3X7PBLCGOTTX","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"35b81f1318fa48e7b81ac9fd4d070ef71a77385c402f78c05c1897f66a100c96","cross_cats_sorted":["stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-22T11:15:07Z","title_canon_sha256":"0621ead406431bf076b3c359eed55de5f726c6f3b68fc4089ef0e44cb642f028"},"schema_version":"1.0","source":{"id":"2410.16901","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2410.16901","created_at":"2026-07-05T09:24:08Z"},{"alias_kind":"arxiv_version","alias_value":"2410.16901v1","created_at":"2026-07-05T09:24:08Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.16901","created_at":"2026-07-05T09:24:08Z"},{"alias_kind":"pith_short_12","alias_value":"IAOFXIVIEKBQ","created_at":"2026-07-05T09:24:08Z"},{"alias_kind":"pith_short_16","alias_value":"IAOFXIVIEKBQME3X","created_at":"2026-07-05T09:24:08Z"},{"alias_kind":"pith_short_8","alias_value":"IAOFXIVI","created_at":"2026-07-05T09:24:08Z"}],"graph_snapshots":[{"event_id":"sha256:8c6beadc56bb22289ca0a5afe8d36d98a72ab0d9dc18ebe881f9e8541015c13c","target":"graph","created_at":"2026-07-05T09:24:08Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2410.16901/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Bayesian deep learning all too often underfits so that the Bayesian prediction is less accurate than a simple point estimate. Uncertainty quantification then comes at the cost of accuracy. For linearized models, the null space of the generalized Gauss-Newton matrix corresponds to parameters that preserve the training predictions of the point estimate. We propose to build Bayesian approximations in this null space, thereby guaranteeing that the Bayesian predictive does not underfit. We suggest a matrix-free algorithm for projecting onto this null space, which scales linearly with the number of ","authors_text":"Hrittik Roy, Marco Miani, S{\\o}ren Hauberg","cross_cats":["stat.ML"],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-22T11:15:07Z","title":"Bayes without Underfitting: Fully Correlated Deep Learning Posteriors via Alternating Projections"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.16901","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:f5af3cefabe68b801bbe0d24683b7441f5f355795a3708a596e5f974f9d39510","target":"record","created_at":"2026-07-05T09:24:08Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"35b81f1318fa48e7b81ac9fd4d070ef71a77385c402f78c05c1897f66a100c96","cross_cats_sorted":["stat.ML"],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-22T11:15:07Z","title_canon_sha256":"0621ead406431bf076b3c359eed55de5f726c6f3b68fc4089ef0e44cb642f028"},"schema_version":"1.0","source":{"id":"2410.16901","kind":"arxiv","version":1}},"canonical_sha256":"401c5ba2a82283061377fbc2b119d39dd3a1e19fcc3ae29ed26f2567cbb5b81b","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"401c5ba2a82283061377fbc2b119d39dd3a1e19fcc3ae29ed26f2567cbb5b81b","first_computed_at":"2026-07-05T09:24:08.381659Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T09:24:08.381659Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"YF39q56rLUG0eGzPKs/QlfBcsQl7JGZLvjriCSZnxA5t+lzjylAG2M6ZSokrcwpvuwTWDYoorg1m8KC5LGdsDA==","signature_status":"signed_v1","signed_at":"2026-07-05T09:24:08.382155Z","signed_message":"canonical_sha256_bytes"},"source_id":"2410.16901","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f5af3cefabe68b801bbe0d24683b7441f5f355795a3708a596e5f974f9d39510","sha256:8c6beadc56bb22289ca0a5afe8d36d98a72ab0d9dc18ebe881f9e8541015c13c"],"state_sha256":"cda8e9e48cc80dc8faef363afcc3550b166193534c37cd92e467f46ebe8b0694"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"2bXUqMtoOE+wL8iznNoSBRiMNdff9ywLAbcUvS6a1aevQekalXVW7HxDZGGWgHOeDkoPEX2Q/HlI19oycMe4Cg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-23T17:37:56.309823Z","bundle_sha256":"8a5aaa5a743dc42411a2e22187b019c1b97ea34cf3a3cb98241ce5cf9657a6fb"}}