{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:YPUCL2YFEPAS2WGZCNSM6N6CMJ","short_pith_number":"pith:YPUCL2YF","schema_version":"1.0","canonical_sha256":"c3e825eb0523c12d58d91364cf37c2625b14a37b13bb6b6fffb42045a19bb762","source":{"kind":"arxiv","id":"2112.14232","version":2},"attestation_state":"computed","paper":{"title":"Constrained Gradient Descent: A Powerful and Principled Evasion Attack Against Neural Networks","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.CR","cs.CV"],"primary_cat":"cs.LG","authors_text":"Keane Lucas, Lujo Bauer, Mahmood Sharif, Michael K. Reiter, Weiran Lin","submitted_at":"2021-12-28T17:36:58Z","abstract_excerpt":"We propose new, more efficient targeted white-box attacks against deep neural networks. Our attacks better align with the attacker's goal: (1) tricking a model to assign higher probability to the target class than to any other class, while (2) staying within an $\\epsilon$-distance of the attacked input. First, we demonstrate a loss function that explicitly encodes (1) and show that Auto-PGD finds more attacks with it. Second, we propose a new attack method, Constrained Gradient Descent (CGD), using a refinement of our loss function that captures both (1) and (2). CGD seeks to satisfy both atta"},"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":"2112.14232","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"cs.LG","submitted_at":"2021-12-28T17:36:58Z","cross_cats_sorted":["cs.CR","cs.CV"],"title_canon_sha256":"3ce1e196169a09c0541ddc49d290cdac11f9d495856a8f325a6371fb0ea7c482","abstract_canon_sha256":"05c6c62977fbe32da68a8cbd08901568e7143f7edd18599399806ca7ff30789a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:33:28.317917Z","signature_b64":"Mi/odAJ/kCuXAaLbv+I4CIeVrHyQOBfgOuT8DScUm68le/IWpDq6EtJkkONUqul1eCtY2KkxjjEy7riN/jLUCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"c3e825eb0523c12d58d91364cf37c2625b14a37b13bb6b6fffb42045a19bb762","last_reissued_at":"2026-07-05T04:33:28.317283Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:33:28.317283Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Constrained Gradient Descent: A Powerful and Principled Evasion Attack Against Neural Networks","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"","cross_cats":["cs.CR","cs.CV"],"primary_cat":"cs.LG","authors_text":"Keane Lucas, Lujo Bauer, Mahmood Sharif, Michael K. Reiter, Weiran Lin","submitted_at":"2021-12-28T17:36:58Z","abstract_excerpt":"We propose new, more efficient targeted white-box attacks against deep neural networks. Our attacks better align with the attacker's goal: (1) tricking a model to assign higher probability to the target class than to any other class, while (2) staying within an $\\epsilon$-distance of the attacked input. First, we demonstrate a loss function that explicitly encodes (1) and show that Auto-PGD finds more attacks with it. Second, we propose a new attack method, Constrained Gradient Descent (CGD), using a refinement of our loss function that captures both (1) and (2). CGD seeks to satisfy both atta"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2112.14232","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/2112.14232/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":"2112.14232","created_at":"2026-07-05T04:33:28.317367+00:00"},{"alias_kind":"arxiv_version","alias_value":"2112.14232v2","created_at":"2026-07-05T04:33:28.317367+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2112.14232","created_at":"2026-07-05T04:33:28.317367+00:00"},{"alias_kind":"pith_short_12","alias_value":"YPUCL2YFEPAS","created_at":"2026-07-05T04:33:28.317367+00:00"},{"alias_kind":"pith_short_16","alias_value":"YPUCL2YFEPAS2WGZ","created_at":"2026-07-05T04:33:28.317367+00:00"},{"alias_kind":"pith_short_8","alias_value":"YPUCL2YF","created_at":"2026-07-05T04:33:28.317367+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/YPUCL2YFEPAS2WGZCNSM6N6CMJ","json":"https://pith.science/pith/YPUCL2YFEPAS2WGZCNSM6N6CMJ.json","graph_json":"https://pith.science/api/pith-number/YPUCL2YFEPAS2WGZCNSM6N6CMJ/graph.json","events_json":"https://pith.science/api/pith-number/YPUCL2YFEPAS2WGZCNSM6N6CMJ/events.json","paper":"https://pith.science/paper/YPUCL2YF"},"agent_actions":{"view_html":"https://pith.science/pith/YPUCL2YFEPAS2WGZCNSM6N6CMJ","download_json":"https://pith.science/pith/YPUCL2YFEPAS2WGZCNSM6N6CMJ.json","view_paper":"https://pith.science/paper/YPUCL2YF","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2112.14232&json=true","fetch_graph":"https://pith.science/api/pith-number/YPUCL2YFEPAS2WGZCNSM6N6CMJ/graph.json","fetch_events":"https://pith.science/api/pith-number/YPUCL2YFEPAS2WGZCNSM6N6CMJ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/YPUCL2YFEPAS2WGZCNSM6N6CMJ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/YPUCL2YFEPAS2WGZCNSM6N6CMJ/action/storage_attestation","attest_author":"https://pith.science/pith/YPUCL2YFEPAS2WGZCNSM6N6CMJ/action/author_attestation","sign_citation":"https://pith.science/pith/YPUCL2YFEPAS2WGZCNSM6N6CMJ/action/citation_signature","submit_replication":"https://pith.science/pith/YPUCL2YFEPAS2WGZCNSM6N6CMJ/action/replication_record"}},"created_at":"2026-07-05T04:33:28.317367+00:00","updated_at":"2026-07-05T04:33:28.317367+00:00"}