{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2017:4MXPQGUXQWWI7TZW57ZR34XF7O","short_pith_number":"pith:4MXPQGUX","canonical_record":{"source":{"id":"1712.02390","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-12-06T19:43:47Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"c669138bba6341bf67678a340b29e4250f2a877eae2a8544b23765b8c783e0c4","abstract_canon_sha256":"f2d28e2bf313ce8e5f321368792dce0866e542dc7691135de252c447a8c5f151"},"schema_version":"1.0"},"canonical_sha256":"e32ef81a9785ac8fcf36eff31df2e5fb9700c0a077db0c279f42805d3b2187d4","source":{"kind":"arxiv","id":"1712.02390","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1712.02390","created_at":"2026-05-18T00:22:39Z"},{"alias_kind":"arxiv_version","alias_value":"1712.02390v2","created_at":"2026-05-18T00:22:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1712.02390","created_at":"2026-05-18T00:22:39Z"},{"alias_kind":"pith_short_12","alias_value":"4MXPQGUXQWWI","created_at":"2026-05-18T12:31:00Z"},{"alias_kind":"pith_short_16","alias_value":"4MXPQGUXQWWI7TZW","created_at":"2026-05-18T12:31:00Z"},{"alias_kind":"pith_short_8","alias_value":"4MXPQGUX","created_at":"2026-05-18T12:31:00Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2017:4MXPQGUXQWWI7TZW57ZR34XF7O","target":"record","payload":{"canonical_record":{"source":{"id":"1712.02390","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-12-06T19:43:47Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"c669138bba6341bf67678a340b29e4250f2a877eae2a8544b23765b8c783e0c4","abstract_canon_sha256":"f2d28e2bf313ce8e5f321368792dce0866e542dc7691135de252c447a8c5f151"},"schema_version":"1.0"},"canonical_sha256":"e32ef81a9785ac8fcf36eff31df2e5fb9700c0a077db0c279f42805d3b2187d4","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:22:39.116448Z","signature_b64":"15NZtlsWXxq3TmGQuh/n17s/9N8X5hLbz90mhuYXVPPenT0uD/VcKXcJ+Ceexk3XzycxZJCYvtivsevNadQLAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"e32ef81a9785ac8fcf36eff31df2e5fb9700c0a077db0c279f42805d3b2187d4","last_reissued_at":"2026-05-18T00:22:39.116079Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:22:39.116079Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1712.02390","source_version":2,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T00:22:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"h5vZ3OVJSe+NQ/0e6VVivMmxg77BzMTA8+affKyW/AlQYEhF2kt32Z0CAvPEkVrckApJRg6ULOb/1TjRc73pDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-01T09:34:41.904813Z"},"content_sha256":"2564902688eedd69ffc9e5ca9069f6a8565a8b2870a985d0a9f8eb39e6c7d430","schema_version":"1.0","event_id":"sha256:2564902688eedd69ffc9e5ca9069f6a8565a8b2870a985d0a9f8eb39e6c7d430"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2017:4MXPQGUXQWWI7TZW57ZR34XF7O","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Noisy Natural Gradient as Variational Inference","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"David Duvenaud, Guodong Zhang, Roger Grosse, Shengyang Sun","submitted_at":"2017-12-06T19:43:47Z","abstract_excerpt":"Variational Bayesian neural nets combine the flexibility of deep learning with Bayesian uncertainty estimation. Unfortunately, there is a tradeoff between cheap but simple variational families (e.g.~fully factorized) or expensive and complicated inference procedures. We show that natural gradient ascent with adaptive weight noise implicitly fits a variational posterior to maximize the evidence lower bound (ELBO). This insight allows us to train full-covariance, fully factorized, or matrix-variate Gaussian variational posteriors using noisy versions of natural gradient, Adam, and K-FAC, respect"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1712.02390","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":""},"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":"2026-05-18T00:22:39Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"YgTCU/I3y8C1ChvL+BQnrikY7oDKc6C81qWqEUF5gEEasSheQ1XEezcrEZUNMKl5euwxcgAwFmZoJbGKluI7BQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-06-01T09:34:41.905178Z"},"content_sha256":"6b910c5c0022f83d1a4f38f6f68d96a63a6b13918fe91210d3c5eaf29a97b036","schema_version":"1.0","event_id":"sha256:6b910c5c0022f83d1a4f38f6f68d96a63a6b13918fe91210d3c5eaf29a97b036"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/4MXPQGUXQWWI7TZW57ZR34XF7O/bundle.json","state_url":"https://pith.science/pith/4MXPQGUXQWWI7TZW57ZR34XF7O/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/4MXPQGUXQWWI7TZW57ZR34XF7O/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":"2026-06-01T09:34:41Z","links":{"resolver":"https://pith.science/pith/4MXPQGUXQWWI7TZW57ZR34XF7O","bundle":"https://pith.science/pith/4MXPQGUXQWWI7TZW57ZR34XF7O/bundle.json","state":"https://pith.science/pith/4MXPQGUXQWWI7TZW57ZR34XF7O/state.json","well_known_bundle":"https://pith.science/.well-known/pith/4MXPQGUXQWWI7TZW57ZR34XF7O/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2017:4MXPQGUXQWWI7TZW57ZR34XF7O","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":"f2d28e2bf313ce8e5f321368792dce0866e542dc7691135de252c447a8c5f151","cross_cats_sorted":["stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-12-06T19:43:47Z","title_canon_sha256":"c669138bba6341bf67678a340b29e4250f2a877eae2a8544b23765b8c783e0c4"},"schema_version":"1.0","source":{"id":"1712.02390","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1712.02390","created_at":"2026-05-18T00:22:39Z"},{"alias_kind":"arxiv_version","alias_value":"1712.02390v2","created_at":"2026-05-18T00:22:39Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1712.02390","created_at":"2026-05-18T00:22:39Z"},{"alias_kind":"pith_short_12","alias_value":"4MXPQGUXQWWI","created_at":"2026-05-18T12:31:00Z"},{"alias_kind":"pith_short_16","alias_value":"4MXPQGUXQWWI7TZW","created_at":"2026-05-18T12:31:00Z"},{"alias_kind":"pith_short_8","alias_value":"4MXPQGUX","created_at":"2026-05-18T12:31:00Z"}],"graph_snapshots":[{"event_id":"sha256:6b910c5c0022f83d1a4f38f6f68d96a63a6b13918fe91210d3c5eaf29a97b036","target":"graph","created_at":"2026-05-18T00:22:39Z","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"},"paper":{"abstract_excerpt":"Variational Bayesian neural nets combine the flexibility of deep learning with Bayesian uncertainty estimation. Unfortunately, there is a tradeoff between cheap but simple variational families (e.g.~fully factorized) or expensive and complicated inference procedures. We show that natural gradient ascent with adaptive weight noise implicitly fits a variational posterior to maximize the evidence lower bound (ELBO). This insight allows us to train full-covariance, fully factorized, or matrix-variate Gaussian variational posteriors using noisy versions of natural gradient, Adam, and K-FAC, respect","authors_text":"David Duvenaud, Guodong Zhang, Roger Grosse, Shengyang Sun","cross_cats":["stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-12-06T19:43:47Z","title":"Noisy Natural Gradient as Variational Inference"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1712.02390","kind":"arxiv","version":2},"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":"sha256:2564902688eedd69ffc9e5ca9069f6a8565a8b2870a985d0a9f8eb39e6c7d430","target":"record","created_at":"2026-05-18T00:22:39Z","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":"f2d28e2bf313ce8e5f321368792dce0866e542dc7691135de252c447a8c5f151","cross_cats_sorted":["stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2017-12-06T19:43:47Z","title_canon_sha256":"c669138bba6341bf67678a340b29e4250f2a877eae2a8544b23765b8c783e0c4"},"schema_version":"1.0","source":{"id":"1712.02390","kind":"arxiv","version":2}},"canonical_sha256":"e32ef81a9785ac8fcf36eff31df2e5fb9700c0a077db0c279f42805d3b2187d4","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"e32ef81a9785ac8fcf36eff31df2e5fb9700c0a077db0c279f42805d3b2187d4","first_computed_at":"2026-05-18T00:22:39.116079Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T00:22:39.116079Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"15NZtlsWXxq3TmGQuh/n17s/9N8X5hLbz90mhuYXVPPenT0uD/VcKXcJ+Ceexk3XzycxZJCYvtivsevNadQLAQ==","signature_status":"signed_v1","signed_at":"2026-05-18T00:22:39.116448Z","signed_message":"canonical_sha256_bytes"},"source_id":"1712.02390","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:2564902688eedd69ffc9e5ca9069f6a8565a8b2870a985d0a9f8eb39e6c7d430","sha256:6b910c5c0022f83d1a4f38f6f68d96a63a6b13918fe91210d3c5eaf29a97b036"],"state_sha256":"a8775fef0f700a3e8bcb93ae8fe1ef8badc0aba160da6136748297f0cfa290b7"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"6SU7qJCNwh+kdjM4qS9lIvAPc1IJCZ1sfFpHxcZRnH80jzXiHMcIA5gNde+XP4XqOQ/HrddfxrefyzehxpvfAw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-06-01T09:34:41.907171Z","bundle_sha256":"35a7ebd26e052124363921bfa26b01e5b4e5ee1ae7991feba0ef6c31e9d5bb41"}}