{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:F36K2MR4KAJXWL6X4IFPEK5OB6","short_pith_number":"pith:F36K2MR4","schema_version":"1.0","canonical_sha256":"2efcad323c50137b2fd7e20af22bae0f8ec46cc5396c75abec2f86969714ad14","source":{"kind":"arxiv","id":"2102.13624","version":2},"attestation_state":"computed","paper":{"title":"What Doesn't Kill You Makes You Robust(er): How to Adversarially Train against Data Poisoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","cs.CV"],"primary_cat":"cs.LG","authors_text":"Gowthami Somepalli, Jonas Geiping, Liam Fowl, Micah Goldblum, Michael Moeller, Tom Goldstein","submitted_at":"2021-02-26T17:54:36Z","abstract_excerpt":"Data poisoning is a threat model in which a malicious actor tampers with training data to manipulate outcomes at inference time. A variety of defenses against this threat model have been proposed, but each suffers from at least one of the following flaws: they are easily overcome by adaptive attacks, they severely reduce testing performance, or they cannot generalize to diverse data poisoning threat models. Adversarial training, and its variants, are currently considered the only empirically strong defense against (inference-time) adversarial attacks. In this work, we extend the adversarial tr"},"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":"2102.13624","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-02-26T17:54:36Z","cross_cats_sorted":["cs.CR","cs.CV"],"title_canon_sha256":"5d9c653e689a7d020d8bcead2d56b2072ca4c3fb23c2c8ca087b2fe23d596740","abstract_canon_sha256":"5dec9e2cd14819f7c4bf3c71858f6d624ee141947e088fce841980cb09017dc7"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:58:01.438051Z","signature_b64":"H8+6KjrXRyaNBB3Lgr4AYZwE38chuh6ADRwuSKz7fj8QcTct9lmicQq5VV8T1QUvjcN0NcZmFTGcuuibchwQCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"2efcad323c50137b2fd7e20af22bae0f8ec46cc5396c75abec2f86969714ad14","last_reissued_at":"2026-07-05T03:58:01.437624Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:58:01.437624Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"What Doesn't Kill You Makes You Robust(er): How to Adversarially Train against Data Poisoning","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CR","cs.CV"],"primary_cat":"cs.LG","authors_text":"Gowthami Somepalli, Jonas Geiping, Liam Fowl, Micah Goldblum, Michael Moeller, Tom Goldstein","submitted_at":"2021-02-26T17:54:36Z","abstract_excerpt":"Data poisoning is a threat model in which a malicious actor tampers with training data to manipulate outcomes at inference time. A variety of defenses against this threat model have been proposed, but each suffers from at least one of the following flaws: they are easily overcome by adaptive attacks, they severely reduce testing performance, or they cannot generalize to diverse data poisoning threat models. Adversarial training, and its variants, are currently considered the only empirically strong defense against (inference-time) adversarial attacks. In this work, we extend the adversarial tr"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2102.13624","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/2102.13624/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":"2102.13624","created_at":"2026-07-05T03:58:01.437680+00:00"},{"alias_kind":"arxiv_version","alias_value":"2102.13624v2","created_at":"2026-07-05T03:58:01.437680+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2102.13624","created_at":"2026-07-05T03:58:01.437680+00:00"},{"alias_kind":"pith_short_12","alias_value":"F36K2MR4KAJX","created_at":"2026-07-05T03:58:01.437680+00:00"},{"alias_kind":"pith_short_16","alias_value":"F36K2MR4KAJXWL6X","created_at":"2026-07-05T03:58:01.437680+00:00"},{"alias_kind":"pith_short_8","alias_value":"F36K2MR4","created_at":"2026-07-05T03:58:01.437680+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2508.20032","citing_title":"Pruning Strategies for Backdoor Defense in LLMs","ref_index":14,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/F36K2MR4KAJXWL6X4IFPEK5OB6","json":"https://pith.science/pith/F36K2MR4KAJXWL6X4IFPEK5OB6.json","graph_json":"https://pith.science/api/pith-number/F36K2MR4KAJXWL6X4IFPEK5OB6/graph.json","events_json":"https://pith.science/api/pith-number/F36K2MR4KAJXWL6X4IFPEK5OB6/events.json","paper":"https://pith.science/paper/F36K2MR4"},"agent_actions":{"view_html":"https://pith.science/pith/F36K2MR4KAJXWL6X4IFPEK5OB6","download_json":"https://pith.science/pith/F36K2MR4KAJXWL6X4IFPEK5OB6.json","view_paper":"https://pith.science/paper/F36K2MR4","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2102.13624&json=true","fetch_graph":"https://pith.science/api/pith-number/F36K2MR4KAJXWL6X4IFPEK5OB6/graph.json","fetch_events":"https://pith.science/api/pith-number/F36K2MR4KAJXWL6X4IFPEK5OB6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/F36K2MR4KAJXWL6X4IFPEK5OB6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/F36K2MR4KAJXWL6X4IFPEK5OB6/action/storage_attestation","attest_author":"https://pith.science/pith/F36K2MR4KAJXWL6X4IFPEK5OB6/action/author_attestation","sign_citation":"https://pith.science/pith/F36K2MR4KAJXWL6X4IFPEK5OB6/action/citation_signature","submit_replication":"https://pith.science/pith/F36K2MR4KAJXWL6X4IFPEK5OB6/action/replication_record"}},"created_at":"2026-07-05T03:58:01.437680+00:00","updated_at":"2026-07-05T03:58:01.437680+00:00"}