{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:DA2W27BX7PKMHLXRHJBMDW5ZMM","short_pith_number":"pith:DA2W27BX","schema_version":"1.0","canonical_sha256":"18356d7c37fbd4c3aef13a42c1dbb96312417299376be2f9d40f7c5723d6bc9d","source":{"kind":"arxiv","id":"2112.08304","version":2},"attestation_state":"computed","paper":{"title":"On the Convergence and Robustness of Adversarial Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bowen Zhou, James Bailey, Jinfeng Yi, Quanquan Gu, Xingjun Ma, Yisen Wang","submitted_at":"2021-12-15T17:54:08Z","abstract_excerpt":"Improving the robustness of deep neural networks (DNNs) to adversarial examples is an important yet challenging problem for secure deep learning. Across existing defense techniques, adversarial training with Projected Gradient Decent (PGD) is amongst the most effective. Adversarial training solves a min-max optimization problem, with the \\textit{inner maximization} generating adversarial examples by maximizing the classification loss, and the \\textit{outer minimization} finding model parameters by minimizing the loss on adversarial examples generated from the inner maximization. A criterion th"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2112.08304","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-12-15T17:54:08Z","cross_cats_sorted":[],"title_canon_sha256":"99dc7cb53f9a1a3945c56bb4cda94ebeac70358edd5b51043e754e2fe25e13df","abstract_canon_sha256":"9f6b9ecbb0398d8a1a1ecfc8008c55374a28afd61c9735ed9db988ae1296f189"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:17:03.803525Z","signature_b64":"R6CQ+t834+fOyH3ORVyoiHueEUDQO7q28I9lekeVZiTJJPKWwQv0zHMHcNh3S9VKXTiSG8IUIa11v38P8swRDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"18356d7c37fbd4c3aef13a42c1dbb96312417299376be2f9d40f7c5723d6bc9d","last_reissued_at":"2026-07-05T04:17:03.803051Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:17:03.803051Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"On the Convergence and Robustness of Adversarial Training","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Bowen Zhou, James Bailey, Jinfeng Yi, Quanquan Gu, Xingjun Ma, Yisen Wang","submitted_at":"2021-12-15T17:54:08Z","abstract_excerpt":"Improving the robustness of deep neural networks (DNNs) to adversarial examples is an important yet challenging problem for secure deep learning. Across existing defense techniques, adversarial training with Projected Gradient Decent (PGD) is amongst the most effective. Adversarial training solves a min-max optimization problem, with the \\textit{inner maximization} generating adversarial examples by maximizing the classification loss, and the \\textit{outer minimization} finding model parameters by minimizing the loss on adversarial examples generated from the inner maximization. A criterion th"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2112.08304","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/2112.08304/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"},"aliases":[{"alias_kind":"arxiv","alias_value":"2112.08304","created_at":"2026-07-05T04:17:03.803102+00:00"},{"alias_kind":"arxiv_version","alias_value":"2112.08304v2","created_at":"2026-07-05T04:17:03.803102+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2112.08304","created_at":"2026-07-05T04:17:03.803102+00:00"},{"alias_kind":"pith_short_12","alias_value":"DA2W27BX7PKM","created_at":"2026-07-05T04:17:03.803102+00:00"},{"alias_kind":"pith_short_16","alias_value":"DA2W27BX7PKMHLXR","created_at":"2026-07-05T04:17:03.803102+00:00"},{"alias_kind":"pith_short_8","alias_value":"DA2W27BX","created_at":"2026-07-05T04:17:03.803102+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2403.09532","citing_title":"Robust SGLD algorithm for solving non-convex distributionally robust optimisation problems","ref_index":129,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/DA2W27BX7PKMHLXRHJBMDW5ZMM","json":"https://pith.science/pith/DA2W27BX7PKMHLXRHJBMDW5ZMM.json","graph_json":"https://pith.science/api/pith-number/DA2W27BX7PKMHLXRHJBMDW5ZMM/graph.json","events_json":"https://pith.science/api/pith-number/DA2W27BX7PKMHLXRHJBMDW5ZMM/events.json","paper":"https://pith.science/paper/DA2W27BX"},"agent_actions":{"view_html":"https://pith.science/pith/DA2W27BX7PKMHLXRHJBMDW5ZMM","download_json":"https://pith.science/pith/DA2W27BX7PKMHLXRHJBMDW5ZMM.json","view_paper":"https://pith.science/paper/DA2W27BX","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2112.08304&json=true","fetch_graph":"https://pith.science/api/pith-number/DA2W27BX7PKMHLXRHJBMDW5ZMM/graph.json","fetch_events":"https://pith.science/api/pith-number/DA2W27BX7PKMHLXRHJBMDW5ZMM/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/DA2W27BX7PKMHLXRHJBMDW5ZMM/action/timestamp_anchor","attest_storage":"https://pith.science/pith/DA2W27BX7PKMHLXRHJBMDW5ZMM/action/storage_attestation","attest_author":"https://pith.science/pith/DA2W27BX7PKMHLXRHJBMDW5ZMM/action/author_attestation","sign_citation":"https://pith.science/pith/DA2W27BX7PKMHLXRHJBMDW5ZMM/action/citation_signature","submit_replication":"https://pith.science/pith/DA2W27BX7PKMHLXRHJBMDW5ZMM/action/replication_record"}},"created_at":"2026-07-05T04:17:03.803102+00:00","updated_at":"2026-07-05T04:17:03.803102+00:00"}