{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:WWCFKU5GPKIBUA4C6TCVZI5GSD","short_pith_number":"pith:WWCFKU5G","schema_version":"1.0","canonical_sha256":"b5845553a67a901a0382f4c55ca3a690c566945c6b89d9650e438c5fa4021def","source":{"kind":"arxiv","id":"2410.06895","version":4},"attestation_state":"computed","paper":{"title":"Average Certified Radius is a Poor Metric for Randomized Smoothing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chenhao Sun, Mark Niklas M\\\"uller, Martin Vechev, Yuhao Mao","submitted_at":"2024-10-09T13:58:41Z","abstract_excerpt":"Randomized smoothing (RS) is popular for providing certified robustness guarantees against adversarial attacks. The average certified radius (ACR) has emerged as a widely used metric for tracking progress in RS. However, in this work, for the first time we show that ACR is a poor metric for evaluating robustness guarantees provided by RS. We theoretically prove not only that a trivial classifier can have arbitrarily large ACR, but also that ACR is extremely sensitive to improvements on easy samples. In addition, the comparison using ACR has a strong dependence on the certification budget. Empi"},"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":"2410.06895","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2024-10-09T13:58:41Z","cross_cats_sorted":[],"title_canon_sha256":"159729d84ff6e8b99c865ec5301bb06fc679dd3d5cd3e660dc84eef3bc827d13","abstract_canon_sha256":"df6223c9330e71d12ee8858c90d849201503f15da4001c1964831e735141d6c2"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T11:16:03.886994Z","signature_b64":"+BHA/TadwZY+2pyFL/qXZ2L7hpHegaCgxfKv6tpNsl+xksX0PsAwj7Qk5CeNfLxtFTZ8yl7ciUiI3rsmOWNtDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b5845553a67a901a0382f4c55ca3a690c566945c6b89d9650e438c5fa4021def","last_reissued_at":"2026-07-05T11:16:03.886519Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T11:16:03.886519Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Average Certified Radius is a Poor Metric for Randomized Smoothing","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chenhao Sun, Mark Niklas M\\\"uller, Martin Vechev, Yuhao Mao","submitted_at":"2024-10-09T13:58:41Z","abstract_excerpt":"Randomized smoothing (RS) is popular for providing certified robustness guarantees against adversarial attacks. The average certified radius (ACR) has emerged as a widely used metric for tracking progress in RS. However, in this work, for the first time we show that ACR is a poor metric for evaluating robustness guarantees provided by RS. We theoretically prove not only that a trivial classifier can have arbitrarily large ACR, but also that ACR is extremely sensitive to improvements on easy samples. In addition, the comparison using ACR has a strong dependence on the certification budget. Empi"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2410.06895","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/2410.06895/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":"2410.06895","created_at":"2026-07-05T11:16:03.886582+00:00"},{"alias_kind":"arxiv_version","alias_value":"2410.06895v4","created_at":"2026-07-05T11:16:03.886582+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2410.06895","created_at":"2026-07-05T11:16:03.886582+00:00"},{"alias_kind":"pith_short_12","alias_value":"WWCFKU5GPKIB","created_at":"2026-07-05T11:16:03.886582+00:00"},{"alias_kind":"pith_short_16","alias_value":"WWCFKU5GPKIBUA4C","created_at":"2026-07-05T11:16:03.886582+00:00"},{"alias_kind":"pith_short_8","alias_value":"WWCFKU5G","created_at":"2026-07-05T11:16:03.886582+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/WWCFKU5GPKIBUA4C6TCVZI5GSD","json":"https://pith.science/pith/WWCFKU5GPKIBUA4C6TCVZI5GSD.json","graph_json":"https://pith.science/api/pith-number/WWCFKU5GPKIBUA4C6TCVZI5GSD/graph.json","events_json":"https://pith.science/api/pith-number/WWCFKU5GPKIBUA4C6TCVZI5GSD/events.json","paper":"https://pith.science/paper/WWCFKU5G"},"agent_actions":{"view_html":"https://pith.science/pith/WWCFKU5GPKIBUA4C6TCVZI5GSD","download_json":"https://pith.science/pith/WWCFKU5GPKIBUA4C6TCVZI5GSD.json","view_paper":"https://pith.science/paper/WWCFKU5G","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2410.06895&json=true","fetch_graph":"https://pith.science/api/pith-number/WWCFKU5GPKIBUA4C6TCVZI5GSD/graph.json","fetch_events":"https://pith.science/api/pith-number/WWCFKU5GPKIBUA4C6TCVZI5GSD/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/WWCFKU5GPKIBUA4C6TCVZI5GSD/action/timestamp_anchor","attest_storage":"https://pith.science/pith/WWCFKU5GPKIBUA4C6TCVZI5GSD/action/storage_attestation","attest_author":"https://pith.science/pith/WWCFKU5GPKIBUA4C6TCVZI5GSD/action/author_attestation","sign_citation":"https://pith.science/pith/WWCFKU5GPKIBUA4C6TCVZI5GSD/action/citation_signature","submit_replication":"https://pith.science/pith/WWCFKU5GPKIBUA4C6TCVZI5GSD/action/replication_record"}},"created_at":"2026-07-05T11:16:03.886582+00:00","updated_at":"2026-07-05T11:16:03.886582+00:00"}