{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2020:BCJIK7FX5BCJW55THZ76PZY2WZ","short_pith_number":"pith:BCJIK7FX","canonical_record":{"source":{"id":"2002.03495","kind":"arxiv","version":14},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-02-10T02:04:49Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"8ea7ea4004d9d4ac3b65ae8e856d0e74566b5b8b156c31911696c111cacc6300","abstract_canon_sha256":"47fbf0227b007908e7c5e3a10ba86225cba2561bc8410aca72c30fc3c381f0ee"},"schema_version":"1.0"},"canonical_sha256":"0892857cb7e8449b77b33e7fe7e71ab67e1b69507ac574c74d42f46b64eb7857","source":{"kind":"arxiv","id":"2002.03495","version":14},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2002.03495","created_at":"2026-07-05T02:07:01Z"},{"alias_kind":"arxiv_version","alias_value":"2002.03495v14","created_at":"2026-07-05T02:07:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2002.03495","created_at":"2026-07-05T02:07:01Z"},{"alias_kind":"pith_short_12","alias_value":"BCJIK7FX5BCJ","created_at":"2026-07-05T02:07:01Z"},{"alias_kind":"pith_short_16","alias_value":"BCJIK7FX5BCJW55T","created_at":"2026-07-05T02:07:01Z"},{"alias_kind":"pith_short_8","alias_value":"BCJIK7FX","created_at":"2026-07-05T02:07:01Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2020:BCJIK7FX5BCJW55THZ76PZY2WZ","target":"record","payload":{"canonical_record":{"source":{"id":"2002.03495","kind":"arxiv","version":14},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-02-10T02:04:49Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"8ea7ea4004d9d4ac3b65ae8e856d0e74566b5b8b156c31911696c111cacc6300","abstract_canon_sha256":"47fbf0227b007908e7c5e3a10ba86225cba2561bc8410aca72c30fc3c381f0ee"},"schema_version":"1.0"},"canonical_sha256":"0892857cb7e8449b77b33e7fe7e71ab67e1b69507ac574c74d42f46b64eb7857","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:07:01.428180Z","signature_b64":"9Dpp3/Uqwpxi41ppntAKYE4+dGI8zM8vlU/M8q5O81sJnq6zFMaKwLk7wAk81sisk8T1BB2GZ4EShTdtzG3yDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0892857cb7e8449b77b33e7fe7e71ab67e1b69507ac574c74d42f46b64eb7857","last_reissued_at":"2026-07-05T02:07:01.427687Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:07:01.427687Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2002.03495","source_version":14,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T02:07:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"j8X4gdVWHNpr/NEHClGqGHs8xbdbUVBgyzulxvHro4kodAllFM3Tqky5HF92YsC3AcV1WGakQLgrKwoVfQr+CQ==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T07:09:37.901578Z"},"content_sha256":"c70cdf0ffc10ae978da2f4f36c6c6519bc033edff41633c5dff41af31cf35371","schema_version":"1.0","event_id":"sha256:c70cdf0ffc10ae978da2f4f36c6c6519bc033edff41633c5dff41af31cf35371"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2020:BCJIK7FX5BCJW55THZ76PZY2WZ","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"A Diffusion Theory For Deep Learning Dynamics: Stochastic Gradient Descent Exponentially Favors Flat Minima","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Issei Sato, Masashi Sugiyama, Zeke Xie","submitted_at":"2020-02-10T02:04:49Z","abstract_excerpt":"Stochastic Gradient Descent (SGD) and its variants are mainstream methods for training deep networks in practice. SGD is known to find a flat minimum that often generalizes well. However, it is mathematically unclear how deep learning can select a flat minimum among so many minima. To answer the question quantitatively, we develop a density diffusion theory (DDT) to reveal how minima selection quantitatively depends on the minima sharpness and the hyperparameters. To the best of our knowledge, we are the first to theoretically and empirically prove that, benefited from the Hessian-dependent co"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2002.03495","kind":"arxiv","version":14},"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/2002.03495/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":"2026-07-05T02:07:01Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"9BhVrR2WeNuZ+7ErDRFL5BeFvnEKY1cDEfiFKhGCc3ou90U6tT1Sgkr2MOY+QwtaJLCafy3+RT42yS/XPDWMCg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T07:09:37.901959Z"},"content_sha256":"49043c1729884d1d358c5842af6ada08e97d986f14ab8fe4d6b48e378a8a388d","schema_version":"1.0","event_id":"sha256:49043c1729884d1d358c5842af6ada08e97d986f14ab8fe4d6b48e378a8a388d"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/BCJIK7FX5BCJW55THZ76PZY2WZ/bundle.json","state_url":"https://pith.science/pith/BCJIK7FX5BCJW55THZ76PZY2WZ/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/BCJIK7FX5BCJW55THZ76PZY2WZ/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":"2026-07-07T07:09:37Z","links":{"resolver":"https://pith.science/pith/BCJIK7FX5BCJW55THZ76PZY2WZ","bundle":"https://pith.science/pith/BCJIK7FX5BCJW55THZ76PZY2WZ/bundle.json","state":"https://pith.science/pith/BCJIK7FX5BCJW55THZ76PZY2WZ/state.json","well_known_bundle":"https://pith.science/.well-known/pith/BCJIK7FX5BCJW55THZ76PZY2WZ/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2020:BCJIK7FX5BCJW55THZ76PZY2WZ","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":"47fbf0227b007908e7c5e3a10ba86225cba2561bc8410aca72c30fc3c381f0ee","cross_cats_sorted":["stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-02-10T02:04:49Z","title_canon_sha256":"8ea7ea4004d9d4ac3b65ae8e856d0e74566b5b8b156c31911696c111cacc6300"},"schema_version":"1.0","source":{"id":"2002.03495","kind":"arxiv","version":14}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2002.03495","created_at":"2026-07-05T02:07:01Z"},{"alias_kind":"arxiv_version","alias_value":"2002.03495v14","created_at":"2026-07-05T02:07:01Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2002.03495","created_at":"2026-07-05T02:07:01Z"},{"alias_kind":"pith_short_12","alias_value":"BCJIK7FX5BCJ","created_at":"2026-07-05T02:07:01Z"},{"alias_kind":"pith_short_16","alias_value":"BCJIK7FX5BCJW55T","created_at":"2026-07-05T02:07:01Z"},{"alias_kind":"pith_short_8","alias_value":"BCJIK7FX","created_at":"2026-07-05T02:07:01Z"}],"graph_snapshots":[{"event_id":"sha256:49043c1729884d1d358c5842af6ada08e97d986f14ab8fe4d6b48e378a8a388d","target":"graph","created_at":"2026-07-05T02:07:01Z","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/2002.03495/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Stochastic Gradient Descent (SGD) and its variants are mainstream methods for training deep networks in practice. SGD is known to find a flat minimum that often generalizes well. However, it is mathematically unclear how deep learning can select a flat minimum among so many minima. To answer the question quantitatively, we develop a density diffusion theory (DDT) to reveal how minima selection quantitatively depends on the minima sharpness and the hyperparameters. To the best of our knowledge, we are the first to theoretically and empirically prove that, benefited from the Hessian-dependent co","authors_text":"Issei Sato, Masashi Sugiyama, Zeke Xie","cross_cats":["stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-02-10T02:04:49Z","title":"A Diffusion Theory For Deep Learning Dynamics: Stochastic Gradient Descent Exponentially Favors Flat Minima"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2002.03495","kind":"arxiv","version":14},"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:c70cdf0ffc10ae978da2f4f36c6c6519bc033edff41633c5dff41af31cf35371","target":"record","created_at":"2026-07-05T02:07:01Z","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":"47fbf0227b007908e7c5e3a10ba86225cba2561bc8410aca72c30fc3c381f0ee","cross_cats_sorted":["stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2020-02-10T02:04:49Z","title_canon_sha256":"8ea7ea4004d9d4ac3b65ae8e856d0e74566b5b8b156c31911696c111cacc6300"},"schema_version":"1.0","source":{"id":"2002.03495","kind":"arxiv","version":14}},"canonical_sha256":"0892857cb7e8449b77b33e7fe7e71ab67e1b69507ac574c74d42f46b64eb7857","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"0892857cb7e8449b77b33e7fe7e71ab67e1b69507ac574c74d42f46b64eb7857","first_computed_at":"2026-07-05T02:07:01.427687Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T02:07:01.427687Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"9Dpp3/Uqwpxi41ppntAKYE4+dGI8zM8vlU/M8q5O81sJnq6zFMaKwLk7wAk81sisk8T1BB2GZ4EShTdtzG3yDg==","signature_status":"signed_v1","signed_at":"2026-07-05T02:07:01.428180Z","signed_message":"canonical_sha256_bytes"},"source_id":"2002.03495","source_kind":"arxiv","source_version":14}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:c70cdf0ffc10ae978da2f4f36c6c6519bc033edff41633c5dff41af31cf35371","sha256:49043c1729884d1d358c5842af6ada08e97d986f14ab8fe4d6b48e378a8a388d"],"state_sha256":"db637f42a77230ce50bfd083e09633ee4c7c150068267aa3ed5428a13d2ff776"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"ElSSSTdsnXDkcdoZCMAvAjfICyYIRLUtztu7dcuv0NlURJPhqf43WsL7ice82qjHMvll3TRgqxULgRNwrJLQBw==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T07:09:37.904401Z","bundle_sha256":"29bd60149432d3e8ee5706038d602dedf2d0e72eb71046f0b4a7862a2d96b00c"}}