{"bundle_type":"pith_open_graph_bundle","bundle_version":"1.0","pith_number":"pith:2018:6CJJSNCDP25NDFAJ3JT426G5E2","short_pith_number":"pith:6CJJSNCD","canonical_record":{"source":{"id":"1805.08736","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-05-22T16:47:46Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3bce3cea3520d77a0d97466e905e06a8bd86e9588db5142641eac8ee9a698962","abstract_canon_sha256":"5e7d4837582a6f0818ef6ddff011a70722c3cc1ef22970ce7bd31dc3558928da"},"schema_version":"1.0"},"canonical_sha256":"f0929934437ebad19409da67cd78dd26872f9f4f5f6acedd36ff16d916be28f5","source":{"kind":"arxiv","id":"1805.08736","version":1},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1805.08736","created_at":"2026-05-18T00:15:25Z"},{"alias_kind":"arxiv_version","alias_value":"1805.08736v1","created_at":"2026-05-18T00:15:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1805.08736","created_at":"2026-05-18T00:15:25Z"},{"alias_kind":"pith_short_12","alias_value":"6CJJSNCDP25N","created_at":"2026-05-18T12:32:08Z"},{"alias_kind":"pith_short_16","alias_value":"6CJJSNCDP25NDFAJ","created_at":"2026-05-18T12:32:08Z"},{"alias_kind":"pith_short_8","alias_value":"6CJJSNCD","created_at":"2026-05-18T12:32:08Z"}],"events":[{"event_type":"record_created","subject_pith_number":"pith:2018:6CJJSNCDP25NDFAJ3JT426G5E2","target":"record","payload":{"canonical_record":{"source":{"id":"1805.08736","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-05-22T16:47:46Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"3bce3cea3520d77a0d97466e905e06a8bd86e9588db5142641eac8ee9a698962","abstract_canon_sha256":"5e7d4837582a6f0818ef6ddff011a70722c3cc1ef22970ce7bd31dc3558928da"},"schema_version":"1.0"},"canonical_sha256":"f0929934437ebad19409da67cd78dd26872f9f4f5f6acedd36ff16d916be28f5","receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-18T00:15:25.217444Z","signature_b64":"ifDRwvHD0qlfWMln5UhSYLz0u3/QxliUkGg3InXknulvsAgWd0X7rWcVJVAJZUmXTEvXa8QtEkoY44bFDnZ6CA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"f0929934437ebad19409da67cd78dd26872f9f4f5f6acedd36ff16d916be28f5","last_reissued_at":"2026-05-18T00:15:25.216934Z","signature_status":"signed_v1","first_computed_at":"2026-05-18T00:15:25.216934Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"source_kind":"arxiv","source_id":"1805.08736","source_version":1,"attestation_state":"computed"},"signer":{"signer_id":"pith.science","signer_type":"pith_registry","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"created_at":"2026-05-18T00:15:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"eiRICru/I4nkWyJDGVS0gqVkeK3FRUq4p4cRMizAHvHzzOGEb/Pe+sXx5spu40LqTmh2emJ3cFhyzD9B72eQCA==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T11:41:58.889904Z"},"content_sha256":"fbf04accd7eff95ab664ed0b49bab36e460181e02aff5bb1aba3251d9784c7cd","schema_version":"1.0","event_id":"sha256:fbf04accd7eff95ab664ed0b49bab36e460181e02aff5bb1aba3251d9784c7cd"},{"event_type":"graph_snapshot","subject_pith_number":"pith:2018:6CJJSNCDP25NDFAJ3JT426G5E2","target":"graph","payload":{"graph_snapshot":{"paper":{"title":"Adversarially Robust Training through Structured Gradient Regularization","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Aurelien Lucchi, Kevin Roth, Sebastian Nowozin, Thomas Hofmann","submitted_at":"2018-05-22T16:47:46Z","abstract_excerpt":"We propose a novel data-dependent structured gradient regularizer to increase the robustness of neural networks vis-a-vis adversarial perturbations. Our regularizer can be derived as a controlled approximation from first principles, leveraging the fundamental link between training with noise and regularization. It adds very little computational overhead during learning and is simple to implement generically in standard deep learning frameworks. Our experiments provide strong evidence that structured gradient regularization can act as an effective first line of defense against attacks based on "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1805.08736","kind":"arxiv","version":1},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"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":"2026-05-18T00:15:25Z","supersedes":[],"prev_event":null,"signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"No4zzpbVk9Re+ojv7qOb/wUgm/w1YSg3s+9Ft/4d07ZK3DWHHjNCoQ9Cdyc1LSnNA+7lAFgHMIxltgHayr44Cg==","signed_message":"open_graph_event_sha256_bytes","signed_at":"2026-05-25T11:41:58.890617Z"},"content_sha256":"502ec80153e30e032b7970736494bc280decf9f52c8861529fa520d2dec699de","schema_version":"1.0","event_id":"sha256:502ec80153e30e032b7970736494bc280decf9f52c8861529fa520d2dec699de"}],"timestamp_proofs":[],"mirror_hints":[{"mirror_type":"https","name":"Pith Resolver","base_url":"https://pith.science","bundle_url":"https://pith.science/pith/6CJJSNCDP25NDFAJ3JT426G5E2/bundle.json","state_url":"https://pith.science/pith/6CJJSNCDP25NDFAJ3JT426G5E2/state.json","well_known_bundle_url":"https://pith.science/.well-known/pith/6CJJSNCDP25NDFAJ3JT426G5E2/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":"2026-05-25T11:41:58Z","links":{"resolver":"https://pith.science/pith/6CJJSNCDP25NDFAJ3JT426G5E2","bundle":"https://pith.science/pith/6CJJSNCDP25NDFAJ3JT426G5E2/bundle.json","state":"https://pith.science/pith/6CJJSNCDP25NDFAJ3JT426G5E2/state.json","well_known_bundle":"https://pith.science/.well-known/pith/6CJJSNCDP25NDFAJ3JT426G5E2/bundle.json"},"state":{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2018:6CJJSNCDP25NDFAJ3JT426G5E2","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":"5e7d4837582a6f0818ef6ddff011a70722c3cc1ef22970ce7bd31dc3558928da","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-05-22T16:47:46Z","title_canon_sha256":"3bce3cea3520d77a0d97466e905e06a8bd86e9588db5142641eac8ee9a698962"},"schema_version":"1.0","source":{"id":"1805.08736","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"1805.08736","created_at":"2026-05-18T00:15:25Z"},{"alias_kind":"arxiv_version","alias_value":"1805.08736v1","created_at":"2026-05-18T00:15:25Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1805.08736","created_at":"2026-05-18T00:15:25Z"},{"alias_kind":"pith_short_12","alias_value":"6CJJSNCDP25N","created_at":"2026-05-18T12:32:08Z"},{"alias_kind":"pith_short_16","alias_value":"6CJJSNCDP25NDFAJ","created_at":"2026-05-18T12:32:08Z"},{"alias_kind":"pith_short_8","alias_value":"6CJJSNCD","created_at":"2026-05-18T12:32:08Z"}],"graph_snapshots":[{"event_id":"sha256:502ec80153e30e032b7970736494bc280decf9f52c8861529fa520d2dec699de","target":"graph","created_at":"2026-05-18T00:15:25Z","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"},"paper":{"abstract_excerpt":"We propose a novel data-dependent structured gradient regularizer to increase the robustness of neural networks vis-a-vis adversarial perturbations. Our regularizer can be derived as a controlled approximation from first principles, leveraging the fundamental link between training with noise and regularization. It adds very little computational overhead during learning and is simple to implement generically in standard deep learning frameworks. Our experiments provide strong evidence that structured gradient regularization can act as an effective first line of defense against attacks based on ","authors_text":"Aurelien Lucchi, Kevin Roth, Sebastian Nowozin, Thomas Hofmann","cross_cats":["cs.LG"],"headline":"","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-05-22T16:47:46Z","title":"Adversarially Robust Training through Structured Gradient Regularization"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1805.08736","kind":"arxiv","version":1},"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:fbf04accd7eff95ab664ed0b49bab36e460181e02aff5bb1aba3251d9784c7cd","target":"record","created_at":"2026-05-18T00:15:25Z","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":"5e7d4837582a6f0818ef6ddff011a70722c3cc1ef22970ce7bd31dc3558928da","cross_cats_sorted":["cs.LG"],"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ML","submitted_at":"2018-05-22T16:47:46Z","title_canon_sha256":"3bce3cea3520d77a0d97466e905e06a8bd86e9588db5142641eac8ee9a698962"},"schema_version":"1.0","source":{"id":"1805.08736","kind":"arxiv","version":1}},"canonical_sha256":"f0929934437ebad19409da67cd78dd26872f9f4f5f6acedd36ff16d916be28f5","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"f0929934437ebad19409da67cd78dd26872f9f4f5f6acedd36ff16d916be28f5","first_computed_at":"2026-05-18T00:15:25.216934Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-05-18T00:15:25.216934Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"ifDRwvHD0qlfWMln5UhSYLz0u3/QxliUkGg3InXknulvsAgWd0X7rWcVJVAJZUmXTEvXa8QtEkoY44bFDnZ6CA==","signature_status":"signed_v1","signed_at":"2026-05-18T00:15:25.217444Z","signed_message":"canonical_sha256_bytes"},"source_id":"1805.08736","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:fbf04accd7eff95ab664ed0b49bab36e460181e02aff5bb1aba3251d9784c7cd","sha256:502ec80153e30e032b7970736494bc280decf9f52c8861529fa520d2dec699de"],"state_sha256":"956771ef01168aec84f574ac0ce636531ad326ec955a06ca14fd564369288ed6"},"bundle_signature":{"signature_status":"signed_v1","algorithm":"ed25519","key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signature_b64":"LNYUb7g9YLQ7e4xmuo+VOKlNnQlrEue5ssuFevpaniwTUz6+sATkq/mGXDYTdML2Aorkzm3rXA2Z0AqOG4BuAg==","signed_message":"bundle_sha256_bytes","signed_at":"2026-05-25T11:41:58.894399Z","bundle_sha256":"1389e3d78ec7cbde12cb0f4d5dab0a12fdd46c658d770bbed3b9bbd1189da95c"}}