{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2024:EGNUSUTRNOMK5KI6GRHIJJXLJE","short_pith_number":"pith:EGNUSUTR","canonical_record":{"source":{"id":"2406.08654","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2024-06-12T21:33:22Z","cross_cats_sorted":["cs.LG","math.OC"],"title_canon_sha256":"9ce31f94f7559c0f8387ad6698df41189a70c910e3ae82c01f6836001f4652dd","abstract_canon_sha256":"210e25253e2478c2e36095460fbb107c28dc449445da28598c26fa0bccd1a4d3"},"schema_version":"1.0"},"canonical_sha256":"219b4952716b98aea91e344e84a6eb490d371f900956275cce6c1316501f5896","source":{"kind":"arxiv","id":"2406.08654","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.08654","created_at":"2026-07-05T08:37:20Z"},{"alias_kind":"arxiv_version","alias_value":"2406.08654v2","created_at":"2026-07-05T08:37:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.08654","created_at":"2026-07-05T08:37:20Z"},{"alias_kind":"pith_short_12","alias_value":"EGNUSUTRNOMK","created_at":"2026-07-05T08:37:20Z"},{"alias_kind":"pith_short_16","alias_value":"EGNUSUTRNOMK5KI6","created_at":"2026-07-05T08:37:20Z"},{"alias_kind":"pith_short_8","alias_value":"EGNUSUTR","created_at":"2026-07-05T08:37:20Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2024:EGNUSUTRNOMK5KI6GRHIJJXLJE","target":"record","payload":{"canonical_record":{"source":{"id":"2406.08654","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2024-06-12T21:33:22Z","cross_cats_sorted":["cs.LG","math.OC"],"title_canon_sha256":"9ce31f94f7559c0f8387ad6698df41189a70c910e3ae82c01f6836001f4652dd","abstract_canon_sha256":"210e25253e2478c2e36095460fbb107c28dc449445da28598c26fa0bccd1a4d3"},"schema_version":"1.0"},"canonical_sha256":"219b4952716b98aea91e344e84a6eb490d371f900956275cce6c1316501f5896","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:37:20.411544Z","signature_b64":"HwitudIIeyuo9bVRPPGdlbHHtsUGWd79xzJ5LoUZKWOe+drsjb2sBiBLZEnR84j3wOJxrHPUDe1b3eYSuUPYAQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"219b4952716b98aea91e344e84a6eb490d371f900956275cce6c1316501f5896","last_reissued_at":"2026-07-05T08:37:20.411129Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:37:20.411129Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2406.08654","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":"2026-07-05T08:37:20Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Yq4JG31KXwfMzuIbVjJJPuHMI39uqaZ6SO46nr8BsZP/8YMoCtmjY8xvdiYuKuM+fD9LTrhFCDA8e+3OHzrABA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-12T07:41:48.242667Z"},"content_sha256":"7d9b3323d7a963d87464db9edf62653c2e33a5acb5b54be156703f59f3b721c1","schema_version":"1.0","event_id":"sha256:7d9b3323d7a963d87464db9edf62653c2e33a5acb5b54be156703f59f3b721c1"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2024:EGNUSUTRNOMK5KI6GRHIJJXLJE","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Large Stepsize Gradient Descent for Non-Homogeneous Two-Layer Networks: Margin Improvement and Fast Optimization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","math.OC"],"primary_cat":"stat.ML","authors_text":"Jingfeng Wu, Michael Lindsey, Peter L. Bartlett, Song Mei, Yuhang Cai","submitted_at":"2024-06-12T21:33:22Z","abstract_excerpt":"The typical training of neural networks using large stepsize gradient descent (GD) under the logistic loss often involves two distinct phases, where the empirical risk oscillates in the first phase but decreases monotonically in the second phase. We investigate this phenomenon in two-layer networks that satisfy a near-homogeneity condition. We show that the second phase begins once the empirical risk falls below a certain threshold, dependent on the stepsize. Additionally, we show that the normalized margin grows nearly monotonically in the second phase, demonstrating an implicit bias of GD in"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.08654","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2406.08654/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":"2026-07-05T08:37:20Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"3eTNz7xufeo3MgrGscDS0PNhE2bD71ph4AMc+FPAy05s1K8HTJ64DIaZx8Hgg4rTqJhWaF+zVwwFSvjTISX/CA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-08-12T07:41:48.243426Z"},"content_sha256":"29d6ac0dd14a5c3a089dc898ecdb574276a57fd7c20180b790b4915236b05dbe","schema_version":"1.0","event_id":"sha256:29d6ac0dd14a5c3a089dc898ecdb574276a57fd7c20180b790b4915236b05dbe"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/EGNUSUTRNOMK5KI6GRHIJJXLJE/bundle.json","state_url":"https://pith.science/pith/EGNUSUTRNOMK5KI6GRHIJJXLJE/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/EGNUSUTRNOMK5KI6GRHIJJXLJE/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":"2026-08-12T07:41:48Z","links":{"resolver":"https://pith.science/pith/EGNUSUTRNOMK5KI6GRHIJJXLJE","bundle":"https://pith.science/pith/EGNUSUTRNOMK5KI6GRHIJJXLJE/bundle.json","state":"https://pith.science/pith/EGNUSUTRNOMK5KI6GRHIJJXLJE/state.json","well_known_bundle":"https://pith.science/.well-known/pith/EGNUSUTRNOMK5KI6GRHIJJXLJE/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2024:EGNUSUTRNOMK5KI6GRHIJJXLJE","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":"210e25253e2478c2e36095460fbb107c28dc449445da28598c26fa0bccd1a4d3","cross_cats_sorted":["cs.LG","math.OC"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2024-06-12T21:33:22Z","title_canon_sha256":"9ce31f94f7559c0f8387ad6698df41189a70c910e3ae82c01f6836001f4652dd"},"schema_version":"1.0","source":{"id":"2406.08654","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2406.08654","created_at":"2026-07-05T08:37:20Z"},{"alias_kind":"arxiv_version","alias_value":"2406.08654v2","created_at":"2026-07-05T08:37:20Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2406.08654","created_at":"2026-07-05T08:37:20Z"},{"alias_kind":"pith_short_12","alias_value":"EGNUSUTRNOMK","created_at":"2026-07-05T08:37:20Z"},{"alias_kind":"pith_short_16","alias_value":"EGNUSUTRNOMK5KI6","created_at":"2026-07-05T08:37:20Z"},{"alias_kind":"pith_short_8","alias_value":"EGNUSUTR","created_at":"2026-07-05T08:37:20Z"}],"graph_snapshots":[{"event_id":"sha256:29d6ac0dd14a5c3a089dc898ecdb574276a57fd7c20180b790b4915236b05dbe","target":"graph","created_at":"2026-07-05T08:37:20Z","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/2406.08654/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"The typical training of neural networks using large stepsize gradient descent (GD) under the logistic loss often involves two distinct phases, where the empirical risk oscillates in the first phase but decreases monotonically in the second phase. We investigate this phenomenon in two-layer networks that satisfy a near-homogeneity condition. We show that the second phase begins once the empirical risk falls below a certain threshold, dependent on the stepsize. Additionally, we show that the normalized margin grows nearly monotonically in the second phase, demonstrating an implicit bias of GD in","authors_text":"Jingfeng Wu, Michael Lindsey, Peter L. Bartlett, Song Mei, Yuhang Cai","cross_cats":["cs.LG","math.OC"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2024-06-12T21:33:22Z","title":"Large Stepsize Gradient Descent for Non-Homogeneous Two-Layer Networks: Margin Improvement and Fast Optimization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2406.08654","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":"sha256:7d9b3323d7a963d87464db9edf62653c2e33a5acb5b54be156703f59f3b721c1","target":"record","created_at":"2026-07-05T08:37:20Z","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":"210e25253e2478c2e36095460fbb107c28dc449445da28598c26fa0bccd1a4d3","cross_cats_sorted":["cs.LG","math.OC"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2024-06-12T21:33:22Z","title_canon_sha256":"9ce31f94f7559c0f8387ad6698df41189a70c910e3ae82c01f6836001f4652dd"},"schema_version":"1.0","source":{"id":"2406.08654","kind":"arxiv","version":2}},"canonical_sha256":"219b4952716b98aea91e344e84a6eb490d371f900956275cce6c1316501f5896","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"219b4952716b98aea91e344e84a6eb490d371f900956275cce6c1316501f5896","first_computed_at":"2026-07-05T08:37:20.411129Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T08:37:20.411129Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"HwitudIIeyuo9bVRPPGdlbHHtsUGWd79xzJ5LoUZKWOe+drsjb2sBiBLZEnR84j3wOJxrHPUDe1b3eYSuUPYAQ==","signature_status":"signed_v1","signed_at":"2026-07-05T08:37:20.411544Z","signed_message":"canonical_sha256_bytes"},"source_id":"2406.08654","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:7d9b3323d7a963d87464db9edf62653c2e33a5acb5b54be156703f59f3b721c1","sha256:29d6ac0dd14a5c3a089dc898ecdb574276a57fd7c20180b790b4915236b05dbe"],"state_sha256":"6143046635b36877b3e18920ee238ed0521d7598fd7219c13290c4d4cf98c3f6"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"kxWVWM35mAMXzJr1mk6ifp0A6zGAZSP6G43qbfmp44nmVk46VAIfTrvk/whK6CXxV+zpUMxpKX0eqeVREDhaCQ==","signed_message":"bundle_sha256_bytes","signed_at":"2026-08-12T07:41:48.318354Z","bundle_sha256":"72cd0e33028073d62deea6ab506ed01de353566253aed9ca014f5357e8c86ca5"}}