{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2021:HPUX5DXDRTJCA47VQDV2WUHAXW","short_pith_number":"pith:HPUX5DXD","canonical_record":{"source":{"id":"2105.08037","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-05-17T17:41:56Z","cross_cats_sorted":["math.OC","math.PR"],"title_canon_sha256":"5df9190b0c0ce34e91e9ca5646dc09a85514650b1064769cec25a76a713842a8","abstract_canon_sha256":"89b71cd501c821fd40e7978a4b84f10947bea8088d77d86684704ccf321c81ef"},"schema_version":"1.0"},"canonical_sha256":"3be97e8ee38cd22073f580ebab50e0bdadbd36b25cac1b529707dc52315a13ac","source":{"kind":"arxiv","id":"2105.08037","version":2},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2105.08037","created_at":"2026-07-05T06:13:14Z"},{"alias_kind":"arxiv_version","alias_value":"2105.08037v2","created_at":"2026-07-05T06:13:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2105.08037","created_at":"2026-07-05T06:13:14Z"},{"alias_kind":"pith_short_12","alias_value":"HPUX5DXDRTJC","created_at":"2026-07-05T06:13:14Z"},{"alias_kind":"pith_short_16","alias_value":"HPUX5DXDRTJCA47V","created_at":"2026-07-05T06:13:14Z"},{"alias_kind":"pith_short_8","alias_value":"HPUX5DXD","created_at":"2026-07-05T06:13:14Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2021:HPUX5DXDRTJCA47VQDV2WUHAXW","target":"record","payload":{"canonical_record":{"source":{"id":"2105.08037","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-05-17T17:41:56Z","cross_cats_sorted":["math.OC","math.PR"],"title_canon_sha256":"5df9190b0c0ce34e91e9ca5646dc09a85514650b1064769cec25a76a713842a8","abstract_canon_sha256":"89b71cd501c821fd40e7978a4b84f10947bea8088d77d86684704ccf321c81ef"},"schema_version":"1.0"},"canonical_sha256":"3be97e8ee38cd22073f580ebab50e0bdadbd36b25cac1b529707dc52315a13ac","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T06:13:14.021052Z","signature_b64":"K1ciWFBYfFjP0dNSOlxYw4o1RGt/pqdW8h4xDMaqUvPDZsxnT+4tcUQ/rDBKSGF+14GPrqskqyL+zhSOzWtdDg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3be97e8ee38cd22073f580ebab50e0bdadbd36b25cac1b529707dc52315a13ac","last_reissued_at":"2026-07-05T06:13:14.020609Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T06:13:14.020609Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"2105.08037","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":"2026-07-05T06:13:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"jRFrt8Viz6pDwguKGZ7QcXIPaKN0nI8o/aze/JOtjAECvrX14UelixUySn2mARByG8K9wgFhT3ZnbtiM0Nz8Dw==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-05T11:55:36.408937Z"},"content_sha256":"9497870d31ad23ef61cc4359b863578c5afe96c5df415649e52e4e27c74cc614","schema_version":"1.0","event_id":"sha256:9497870d31ad23ef61cc4359b863578c5afe96c5df415649e52e4e27c74cc614"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2021:HPUX5DXDRTJCA47VQDV2WUHAXW","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Adversarial Training for Gradient Descent: Analysis Through its Continuous-time Approximation","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["math.OC","math.PR"],"primary_cat":"cs.LG","authors_text":"Haotian Gu, Xin Guo, Xinyu Li","submitted_at":"2021-05-17T17:41:56Z","abstract_excerpt":"Adversarial training has gained great popularity as one of the most effective defenses for deep neural network and more generally for gradient-based machine learning models against adversarial perturbations on data points. This paper establishes a continuous-time approximation for the mini-max game of adversarial training. This approximation approach allows for precise and analytical comparisons between stochastic gradient descent and its adversarial training counterpart; and confirms theoretically the robustness of adversarial training from a new gradient-flow viewpoint. The analysis is then "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2105.08037","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/2105.08037/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":"2026-07-05T06:13:14Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"RVUAyr/jNk+sW6UT/3m4IPIr6U34ybIAxcDqyacGdI97PYlphttsIcbo3wfZht6frvVJNWtzo1WIhcemX/WODg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-07-05T11:55:36.409539Z"},"content_sha256":"5954233e28ee5762f0add17c7f2bd56b9c6ab75d091dcfcff8a1093ac9086465","schema_version":"1.0","event_id":"sha256:5954233e28ee5762f0add17c7f2bd56b9c6ab75d091dcfcff8a1093ac9086465"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/HPUX5DXDRTJCA47VQDV2WUHAXW/bundle.json","state_url":"https://pith.science/pith/HPUX5DXDRTJCA47VQDV2WUHAXW/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/HPUX5DXDRTJCA47VQDV2WUHAXW/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":"2026-07-05T11:55:36Z","links":{"resolver":"https://pith.science/pith/HPUX5DXDRTJCA47VQDV2WUHAXW","bundle":"https://pith.science/pith/HPUX5DXDRTJCA47VQDV2WUHAXW/bundle.json","state":"https://pith.science/pith/HPUX5DXDRTJCA47VQDV2WUHAXW/state.json","well_known_bundle":"https://pith.science/.well-known/pith/HPUX5DXDRTJCA47VQDV2WUHAXW/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2021:HPUX5DXDRTJCA47VQDV2WUHAXW","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":"89b71cd501c821fd40e7978a4b84f10947bea8088d77d86684704ccf321c81ef","cross_cats_sorted":["math.OC","math.PR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-05-17T17:41:56Z","title_canon_sha256":"5df9190b0c0ce34e91e9ca5646dc09a85514650b1064769cec25a76a713842a8"},"schema_version":"1.0","source":{"id":"2105.08037","kind":"arxiv","version":2}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2105.08037","created_at":"2026-07-05T06:13:14Z"},{"alias_kind":"arxiv_version","alias_value":"2105.08037v2","created_at":"2026-07-05T06:13:14Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2105.08037","created_at":"2026-07-05T06:13:14Z"},{"alias_kind":"pith_short_12","alias_value":"HPUX5DXDRTJC","created_at":"2026-07-05T06:13:14Z"},{"alias_kind":"pith_short_16","alias_value":"HPUX5DXDRTJCA47V","created_at":"2026-07-05T06:13:14Z"},{"alias_kind":"pith_short_8","alias_value":"HPUX5DXD","created_at":"2026-07-05T06:13:14Z"}],"graph_snapshots":[{"event_id":"sha256:5954233e28ee5762f0add17c7f2bd56b9c6ab75d091dcfcff8a1093ac9086465","target":"graph","created_at":"2026-07-05T06:13:14Z","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/2105.08037/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Adversarial training has gained great popularity as one of the most effective defenses for deep neural network and more generally for gradient-based machine learning models against adversarial perturbations on data points. This paper establishes a continuous-time approximation for the mini-max game of adversarial training. This approximation approach allows for precise and analytical comparisons between stochastic gradient descent and its adversarial training counterpart; and confirms theoretically the robustness of adversarial training from a new gradient-flow viewpoint. The analysis is then ","authors_text":"Haotian Gu, Xin Guo, Xinyu Li","cross_cats":["math.OC","math.PR"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-05-17T17:41:56Z","title":"Adversarial Training for Gradient Descent: Analysis Through its Continuous-time Approximation"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2105.08037","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":"sha256:9497870d31ad23ef61cc4359b863578c5afe96c5df415649e52e4e27c74cc614","target":"record","created_at":"2026-07-05T06:13:14Z","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":"89b71cd501c821fd40e7978a4b84f10947bea8088d77d86684704ccf321c81ef","cross_cats_sorted":["math.OC","math.PR"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-05-17T17:41:56Z","title_canon_sha256":"5df9190b0c0ce34e91e9ca5646dc09a85514650b1064769cec25a76a713842a8"},"schema_version":"1.0","source":{"id":"2105.08037","kind":"arxiv","version":2}},"canonical_sha256":"3be97e8ee38cd22073f580ebab50e0bdadbd36b25cac1b529707dc52315a13ac","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3be97e8ee38cd22073f580ebab50e0bdadbd36b25cac1b529707dc52315a13ac","first_computed_at":"2026-07-05T06:13:14.020609Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-07-05T06:13:14.020609Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"K1ciWFBYfFjP0dNSOlxYw4o1RGt/pqdW8h4xDMaqUvPDZsxnT+4tcUQ/rDBKSGF+14GPrqskqyL+zhSOzWtdDg==","signature_status":"signed_v1","signed_at":"2026-07-05T06:13:14.021052Z","signed_message":"canonical_sha256_bytes"},"source_id":"2105.08037","source_kind":"arxiv","source_version":2}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:9497870d31ad23ef61cc4359b863578c5afe96c5df415649e52e4e27c74cc614","sha256:5954233e28ee5762f0add17c7f2bd56b9c6ab75d091dcfcff8a1093ac9086465"],"state_sha256":"c676ed3e2ab6cd91326896a5e615c9d70d21d7fd37759bb642e0ab2216380320"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"Z+MG0HpBaObx5DzG3QeAC1IltickFe/hXkmtwaJtWmf1TBYnoab4zx/vuy/MgmuzPYi7ZdvD2j+PWLFAFd2NCg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-07-05T11:55:36.412241Z","bundle_sha256":"efa8a0b2d22d5ddcfe78cdcf36210f4958441817a85395b0b86f6f31ac7a290d"}}