{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:JBKTJBCUHFGFYQZDZKHY3R4CZ6","short_pith_number":"pith:JBKTJBCU","schema_version":"1.0","canonical_sha256":"4855348454394c5c4323ca8f8dc782cfa510b1d391d6637ef26f3c84fb402b1d","source":{"kind":"arxiv","id":"2101.09306","version":4},"attestation_state":"computed","paper":{"title":"Towards Optimal Branching of Linear and Semidefinite Relaxations for Neural Network Robustness Certification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC"],"primary_cat":"cs.LG","authors_text":"Brendon G. Anderson, Jingqi Li, Somayeh Sojoudi, Ziye Ma","submitted_at":"2021-01-22T19:36:40Z","abstract_excerpt":"In this paper, we study certifying the robustness of ReLU neural networks against adversarial input perturbations. To diminish the relaxation error suffered by the popular linear programming (LP) and semidefinite programming (SDP) certification methods, we take a branch-and-bound approach to propose partitioning the input uncertainty set and solving the relaxations on each part separately. We show that this approach reduces relaxation error, and that the error is eliminated entirely upon performing an LP relaxation with a partition intelligently designed to exploit the nature of the ReLU activ"},"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":"2101.09306","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2021-01-22T19:36:40Z","cross_cats_sorted":["math.OC"],"title_canon_sha256":"7b0e2bd4b8c470dd61e5585124435793ff22f43c857e0366af86ebf67ab8c261","abstract_canon_sha256":"20dc3e6c78aa521f99a6f65aeabf7d7eadccd4a9a3a9b3c12a29beba272b32df"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:00:56.941243Z","signature_b64":"Xjhl43AmI16yLCcfZsmD1Sj5gnpivi6M+qHedUCezXW4IH9oWvOfmfXl3pd9k8ySCswCX1Au61SjHvSVFmXHDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4855348454394c5c4323ca8f8dc782cfa510b1d391d6637ef26f3c84fb402b1d","last_reissued_at":"2026-07-05T11:00:56.940691Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:00:56.940691Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards Optimal Branching of Linear and Semidefinite Relaxations for Neural Network Robustness Certification","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["math.OC"],"primary_cat":"cs.LG","authors_text":"Brendon G. Anderson, Jingqi Li, Somayeh Sojoudi, Ziye Ma","submitted_at":"2021-01-22T19:36:40Z","abstract_excerpt":"In this paper, we study certifying the robustness of ReLU neural networks against adversarial input perturbations. To diminish the relaxation error suffered by the popular linear programming (LP) and semidefinite programming (SDP) certification methods, we take a branch-and-bound approach to propose partitioning the input uncertainty set and solving the relaxations on each part separately. We show that this approach reduces relaxation error, and that the error is eliminated entirely upon performing an LP relaxation with a partition intelligently designed to exploit the nature of the ReLU activ"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2101.09306","kind":"arxiv","version":4},"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/2101.09306/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":"2101.09306","created_at":"2026-07-05T11:00:56.940743+00:00"},{"alias_kind":"arxiv_version","alias_value":"2101.09306v4","created_at":"2026-07-05T11:00:56.940743+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2101.09306","created_at":"2026-07-05T11:00:56.940743+00:00"},{"alias_kind":"pith_short_12","alias_value":"JBKTJBCUHFGF","created_at":"2026-07-05T11:00:56.940743+00:00"},{"alias_kind":"pith_short_16","alias_value":"JBKTJBCUHFGFYQZD","created_at":"2026-07-05T11:00:56.940743+00:00"},{"alias_kind":"pith_short_8","alias_value":"JBKTJBCU","created_at":"2026-07-05T11:00:56.940743+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.06665","citing_title":"SDP-CROWN: Efficient Bound Propagation for Neural Network Verification with Tightness of Semidefinite Programming","ref_index":1,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JBKTJBCUHFGFYQZDZKHY3R4CZ6","json":"https://pith.science/pith/JBKTJBCUHFGFYQZDZKHY3R4CZ6.json","graph_json":"https://pith.science/api/pith-number/JBKTJBCUHFGFYQZDZKHY3R4CZ6/graph.json","events_json":"https://pith.science/api/pith-number/JBKTJBCUHFGFYQZDZKHY3R4CZ6/events.json","paper":"https://pith.science/paper/JBKTJBCU"},"agent_actions":{"view_html":"https://pith.science/pith/JBKTJBCUHFGFYQZDZKHY3R4CZ6","download_json":"https://pith.science/pith/JBKTJBCUHFGFYQZDZKHY3R4CZ6.json","view_paper":"https://pith.science/paper/JBKTJBCU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2101.09306&json=true","fetch_graph":"https://pith.science/api/pith-number/JBKTJBCUHFGFYQZDZKHY3R4CZ6/graph.json","fetch_events":"https://pith.science/api/pith-number/JBKTJBCUHFGFYQZDZKHY3R4CZ6/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JBKTJBCUHFGFYQZDZKHY3R4CZ6/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JBKTJBCUHFGFYQZDZKHY3R4CZ6/action/storage_attestation","attest_author":"https://pith.science/pith/JBKTJBCUHFGFYQZDZKHY3R4CZ6/action/author_attestation","sign_citation":"https://pith.science/pith/JBKTJBCUHFGFYQZDZKHY3R4CZ6/action/citation_signature","submit_replication":"https://pith.science/pith/JBKTJBCUHFGFYQZDZKHY3R4CZ6/action/replication_record"}},"created_at":"2026-07-05T11:00:56.940743+00:00","updated_at":"2026-07-05T11:00:56.940743+00:00"}