{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:KCIS5L36RWRNMW63PL7NWD7RKC","short_pith_number":"pith:KCIS5L36","schema_version":"1.0","canonical_sha256":"50912eaf7e8da2d65bdb7afedb0ff150a578248ee7841df24c9bda6e472fe7da","source":{"kind":"arxiv","id":"2106.02867","version":1},"attestation_state":"computed","paper":{"title":"Ensemble Defense with Data Diversity: Weak Correlation Implies Strong Robustness","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Aimin Zhou, Bai Xue, Cheng-Chao Huang, Hanwei Zhang, Lijun Zhang, Pengfei Yang, Renjue Li","submitted_at":"2021-06-05T10:56:48Z","abstract_excerpt":"In this paper, we propose a framework of filter-based ensemble of deep neuralnetworks (DNNs) to defend against adversarial attacks. The framework builds an ensemble of sub-models -- DNNs with differentiated preprocessing filters. From the theoretical perspective of DNN robustness, we argue that under the assumption of high quality of the filters, the weaker the correlations of the sensitivity of the filters are, the more robust the ensemble model tends to be, and this is corroborated by the experiments of transfer-based attacks. Correspondingly, we propose a principle that chooses the specific"},"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":"2106.02867","kind":"arxiv","version":1},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2021-06-05T10:56:48Z","cross_cats_sorted":[],"title_canon_sha256":"3c41b5595015562734a652b4ab5f1fe6c9cb998a35caa4afa9e367349bc1a9bd","abstract_canon_sha256":"ab517d54fd01ae5a41926d6c623449d5960901b8e90f4aad2bd1444a6cc6fcda"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T02:46:31.127205Z","signature_b64":"oxOBv+k82BHcRWISc/rQLU6ImXqIQEKNBXh5ycZe0oqzn9nagZt5cjAHF+t4eykqTAAZS1rlNrogKUxY7BHVBQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"50912eaf7e8da2d65bdb7afedb0ff150a578248ee7841df24c9bda6e472fe7da","last_reissued_at":"2026-07-05T02:46:31.126807Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T02:46:31.126807Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Ensemble Defense with Data Diversity: Weak Correlation Implies Strong Robustness","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Aimin Zhou, Bai Xue, Cheng-Chao Huang, Hanwei Zhang, Lijun Zhang, Pengfei Yang, Renjue Li","submitted_at":"2021-06-05T10:56:48Z","abstract_excerpt":"In this paper, we propose a framework of filter-based ensemble of deep neuralnetworks (DNNs) to defend against adversarial attacks. The framework builds an ensemble of sub-models -- DNNs with differentiated preprocessing filters. From the theoretical perspective of DNN robustness, we argue that under the assumption of high quality of the filters, the weaker the correlations of the sensitivity of the filters are, the more robust the ensemble model tends to be, and this is corroborated by the experiments of transfer-based attacks. Correspondingly, we propose a principle that chooses the specific"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2106.02867","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":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2106.02867/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":"2106.02867","created_at":"2026-07-05T02:46:31.126866+00:00"},{"alias_kind":"arxiv_version","alias_value":"2106.02867v1","created_at":"2026-07-05T02:46:31.126866+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2106.02867","created_at":"2026-07-05T02:46:31.126866+00:00"},{"alias_kind":"pith_short_12","alias_value":"KCIS5L36RWRN","created_at":"2026-07-05T02:46:31.126866+00:00"},{"alias_kind":"pith_short_16","alias_value":"KCIS5L36RWRNMW63","created_at":"2026-07-05T02:46:31.126866+00:00"},{"alias_kind":"pith_short_8","alias_value":"KCIS5L36","created_at":"2026-07-05T02:46:31.126866+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/KCIS5L36RWRNMW63PL7NWD7RKC","json":"https://pith.science/pith/KCIS5L36RWRNMW63PL7NWD7RKC.json","graph_json":"https://pith.science/api/pith-number/KCIS5L36RWRNMW63PL7NWD7RKC/graph.json","events_json":"https://pith.science/api/pith-number/KCIS5L36RWRNMW63PL7NWD7RKC/events.json","paper":"https://pith.science/paper/KCIS5L36"},"agent_actions":{"view_html":"https://pith.science/pith/KCIS5L36RWRNMW63PL7NWD7RKC","download_json":"https://pith.science/pith/KCIS5L36RWRNMW63PL7NWD7RKC.json","view_paper":"https://pith.science/paper/KCIS5L36","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2106.02867&json=true","fetch_graph":"https://pith.science/api/pith-number/KCIS5L36RWRNMW63PL7NWD7RKC/graph.json","fetch_events":"https://pith.science/api/pith-number/KCIS5L36RWRNMW63PL7NWD7RKC/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KCIS5L36RWRNMW63PL7NWD7RKC/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KCIS5L36RWRNMW63PL7NWD7RKC/action/storage_attestation","attest_author":"https://pith.science/pith/KCIS5L36RWRNMW63PL7NWD7RKC/action/author_attestation","sign_citation":"https://pith.science/pith/KCIS5L36RWRNMW63PL7NWD7RKC/action/citation_signature","submit_replication":"https://pith.science/pith/KCIS5L36RWRNMW63PL7NWD7RKC/action/replication_record"}},"created_at":"2026-07-05T02:46:31.126866+00:00","updated_at":"2026-07-05T02:46:31.126866+00:00"}