{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2018:N2HG5O3YGO2C5BS73SNH7EOVPX","short_pith_number":"pith:N2HG5O3Y","canonical_record":{"source":{"id":"1802.05155","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-02-14T15:26:59Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"cbf0623b1397bc4c3188c3fbd431893d7b1c52758660eb93cd77a6e333bd32f6","abstract_canon_sha256":"1e831bd5922b58fec08db7d3b3542bb7f538c46642c35933b93ca98d924992c7"},"schema_version":"1.0"},"canonical_sha256":"6e8e6ebb7833b42e865fdc9a7f91d57dde4c7d088fc68355b1c865a11b42e4d6","source":{"kind":"arxiv","id":"1802.05155","version":5},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1802.05155","created_at":"2026-07-05T02:20:51Z"},{"alias_kind":"arxiv_version","alias_value":"1802.05155v5","created_at":"2026-07-05T02:20:51Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1802.05155","created_at":"2026-07-05T02:20:51Z"},{"alias_kind":"pith_short_12","alias_value":"N2HG5O3YGO2C","created_at":"2026-07-05T02:20:51Z"},{"alias_kind":"pith_short_16","alias_value":"N2HG5O3YGO2C5BS7","created_at":"2026-07-05T02:20:51Z"},{"alias_kind":"pith_short_8","alias_value":"N2HG5O3Y","created_at":"2026-07-05T02:20:51Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2018:N2HG5O3YGO2C5BS73SNH7EOVPX","target":"record","payload":{"canonical_record":{"source":{"id":"1802.05155","kind":"arxiv","version":5},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-02-14T15:26:59Z","cross_cats_sorted":["math.OC","stat.ML"],"title_canon_sha256":"cbf0623b1397bc4c3188c3fbd431893d7b1c52758660eb93cd77a6e333bd32f6","abstract_canon_sha256":"1e831bd5922b58fec08db7d3b3542bb7f538c46642c35933b93ca98d924992c7"},"schema_version":"1.0"},"canonical_sha256":"6e8e6ebb7833b42e865fdc9a7f91d57dde4c7d088fc68355b1c865a11b42e4d6","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:20:51.091598Z","signature_b64":"IQtoN+DpGLxTWd6xUnf9CYys795k138HU7tMQg2voMWmT2uGwdc5uH4q1bEXFZlQIszdd4cMP2arMfNaJy+YCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"6e8e6ebb7833b42e865fdc9a7f91d57dde4c7d088fc68355b1c865a11b42e4d6","last_reissued_at":"2026-07-05T02:20:51.091114Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:20:51.091114Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1802.05155","source_version":5,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-07-05T02:20:51Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"QxAttoYid8zxGl0tCtiUsLkhzUg+oMF862lefKEAcP5TnllyJuvNPNyzBtir6erBdNpWqVnSwNnvB9DcyECnAg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T11:14:42.445102Z"},"content_sha256":"e34f02d2c2b9b5e6c60e2b0f243a9980ab86626d6e4916f1ac4f316be94e9d8d","schema_version":"1.0","event_id":"sha256:e34f02d2c2b9b5e6c60e2b0f243a9980ab86626d6e4916f1ac4f316be94e9d8d"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2018:N2HG5O3YGO2C5BS73SNH7EOVPX","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"A Diffusion Approximation Theory of Momentum SGD in Nonconvex Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","stat.ML"],"primary_cat":"cs.LG","authors_text":"Enlu Zhou, Tianyi Liu, Tuo Zhao, Zhehui Chen","submitted_at":"2018-02-14T15:26:59Z","abstract_excerpt":"Momentum Stochastic Gradient Descent (MSGD) algorithm has been widely applied to many nonconvex optimization problems in machine learning, e.g., training deep neural networks, variational Bayesian inference, and etc. Despite its empirical success, there is still a lack of theoretical understanding of convergence properties of MSGD. To fill this gap, we propose to analyze the algorithmic behavior of MSGD by diffusion approximations for nonconvex optimization problems with strict saddle points and isolated local optima. Our study shows that the momentum helps escape from saddle points, but hurts"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1802.05155","kind":"arxiv","version":5},"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/1802.05155/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":"2026-07-05T02:20:51Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"NT2bPelhv5SYGIv/R8cJtldKbV0kLala6ytD27H2LEB+2CcYHLovAUtkWeor3XbtmIFJ4umaJky5ppN5BVNPDA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-07T11:14:42.445535Z"},"content_sha256":"6d3b3cff0ed3f1c5ad26f4e1ecdac260efc1b648262afc7ba490235a2ecc13fa","schema_version":"1.0","event_id":"sha256:6d3b3cff0ed3f1c5ad26f4e1ecdac260efc1b648262afc7ba490235a2ecc13fa"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/N2HG5O3YGO2C5BS73SNH7EOVPX/bundle.json","state_url":"https://pith.science/pith/N2HG5O3YGO2C5BS73SNH7EOVPX/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/N2HG5O3YGO2C5BS73SNH7EOVPX/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":"2026-07-07T11:14:42Z","links":{"resolver":"https://pith.science/pith/N2HG5O3YGO2C5BS73SNH7EOVPX","bundle":"https://pith.science/pith/N2HG5O3YGO2C5BS73SNH7EOVPX/bundle.json","state":"https://pith.science/pith/N2HG5O3YGO2C5BS73SNH7EOVPX/state.json","well_known_bundle":"https://pith.science/.well-known/pith/N2HG5O3YGO2C5BS73SNH7EOVPX/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2018:N2HG5O3YGO2C5BS73SNH7EOVPX","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":"1e831bd5922b58fec08db7d3b3542bb7f538c46642c35933b93ca98d924992c7","cross_cats_sorted":["math.OC","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-02-14T15:26:59Z","title_canon_sha256":"cbf0623b1397bc4c3188c3fbd431893d7b1c52758660eb93cd77a6e333bd32f6"},"schema_version":"1.0","source":{"id":"1802.05155","kind":"arxiv","version":5}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1802.05155","created_at":"2026-07-05T02:20:51Z"},{"alias_kind":"arxiv_version","alias_value":"1802.05155v5","created_at":"2026-07-05T02:20:51Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1802.05155","created_at":"2026-07-05T02:20:51Z"},{"alias_kind":"pith_short_12","alias_value":"N2HG5O3YGO2C","created_at":"2026-07-05T02:20:51Z"},{"alias_kind":"pith_short_16","alias_value":"N2HG5O3YGO2C5BS7","created_at":"2026-07-05T02:20:51Z"},{"alias_kind":"pith_short_8","alias_value":"N2HG5O3Y","created_at":"2026-07-05T02:20:51Z"}],"graph_snapshots":[{"event_id":"sha256:6d3b3cff0ed3f1c5ad26f4e1ecdac260efc1b648262afc7ba490235a2ecc13fa","target":"graph","created_at":"2026-07-05T02:20:51Z","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/1802.05155/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Momentum Stochastic Gradient Descent (MSGD) algorithm has been widely applied to many nonconvex optimization problems in machine learning, e.g., training deep neural networks, variational Bayesian inference, and etc. Despite its empirical success, there is still a lack of theoretical understanding of convergence properties of MSGD. To fill this gap, we propose to analyze the algorithmic behavior of MSGD by diffusion approximations for nonconvex optimization problems with strict saddle points and isolated local optima. Our study shows that the momentum helps escape from saddle points, but hurts","authors_text":"Enlu Zhou, Tianyi Liu, Tuo Zhao, Zhehui Chen","cross_cats":["math.OC","stat.ML"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-02-14T15:26:59Z","title":"A Diffusion Approximation Theory of Momentum SGD in Nonconvex Optimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1802.05155","kind":"arxiv","version":5},"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:e34f02d2c2b9b5e6c60e2b0f243a9980ab86626d6e4916f1ac4f316be94e9d8d","target":"record","created_at":"2026-07-05T02:20:51Z","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":"1e831bd5922b58fec08db7d3b3542bb7f538c46642c35933b93ca98d924992c7","cross_cats_sorted":["math.OC","stat.ML"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-02-14T15:26:59Z","title_canon_sha256":"cbf0623b1397bc4c3188c3fbd431893d7b1c52758660eb93cd77a6e333bd32f6"},"schema_version":"1.0","source":{"id":"1802.05155","kind":"arxiv","version":5}},"canonical_sha256":"6e8e6ebb7833b42e865fdc9a7f91d57dde4c7d088fc68355b1c865a11b42e4d6","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"6e8e6ebb7833b42e865fdc9a7f91d57dde4c7d088fc68355b1c865a11b42e4d6","first_computed_at":"2026-07-05T02:20:51.091114Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T02:20:51.091114Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"IQtoN+DpGLxTWd6xUnf9CYys795k138HU7tMQg2voMWmT2uGwdc5uH4q1bEXFZlQIszdd4cMP2arMfNaJy+YCQ==","signature_status":"signed_v1","signed_at":"2026-07-05T02:20:51.091598Z","signed_message":"canonical_sha256_bytes"},"source_id":"1802.05155","source_kind":"arxiv","source_version":5}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:e34f02d2c2b9b5e6c60e2b0f243a9980ab86626d6e4916f1ac4f316be94e9d8d","sha256:6d3b3cff0ed3f1c5ad26f4e1ecdac260efc1b648262afc7ba490235a2ecc13fa"],"state_sha256":"872663e3c8ad63f8aad0501ea8a08287a7ee8a83b438fff4532dd6774906b9f1"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"+mCnETuv+VIgE4sH15Nlwwv09iCZdVIKQQDWXzM4RrWlOOvoreoTfyQz9N/aZW158EiLEdg9rObdK5HruNaBDA==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-07T11:14:42.447457Z","bundle_sha256":"88bc622499d0f3e292481f60e6178b24697ce7b3f8b396e85759f6f8a8272486"}}