{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:RONP62W2ZYUEJMOL7PNAKGDGMZ","short_pith_number":"pith:RONP62W2","schema_version":"1.0","canonical_sha256":"8b9aff6adace2844b1cbfbda0518666642952bf37d23c216129e389ff3bb6ecb","source":{"kind":"arxiv","id":"2411.14946","version":1},"attestation_state":"computed","paper":{"title":"Reliable Evaluation of Attribution Maps in CNNs: A Perturbation-Based Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Henrike Stephani, Janis Keuper, Lars Nieradzik","submitted_at":"2024-11-22T13:57:56Z","abstract_excerpt":"In this paper, we present an approach for evaluating attribution maps, which play a central role in interpreting the predictions of convolutional neural networks (CNNs). We show that the widely used insertion/deletion metrics are susceptible to distribution shifts that affect the reliability of the ranking. Our method proposes to replace pixel modifications with adversarial perturbations, which provides a more robust evaluation framework. By using smoothness and monotonicity measures, we illustrate the effectiveness of our approach in correcting distribution shifts. In addition, we conduct the"},"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":"2411.14946","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.CV","submitted_at":"2024-11-22T13:57:56Z","cross_cats_sorted":["cs.AI","cs.LG"],"title_canon_sha256":"cd3d83f12c4e165d6129b547591fc2c461d120df73748d55b495b5dc9f06db14","abstract_canon_sha256":"7deffd481a8e18f9bd33360caae4bc134e743e4b9a6cac0532eb2e1b167b2bf5"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T09:39:14.038002Z","signature_b64":"/yWs4l7PJf6ZBLlw8b3Q5h/ztHhNjykgzEFkpc0Hgb1tQFC5KLvCwM4DWTBsXxsHT/c9uh1/omj4nFGShBQUCQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"8b9aff6adace2844b1cbfbda0518666642952bf37d23c216129e389ff3bb6ecb","last_reissued_at":"2026-07-05T09:39:14.037590Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T09:39:14.037590Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Reliable Evaluation of Attribution Maps in CNNs: A Perturbation-Based Approach","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":["cs.AI","cs.LG"],"primary_cat":"cs.CV","authors_text":"Henrike Stephani, Janis Keuper, Lars Nieradzik","submitted_at":"2024-11-22T13:57:56Z","abstract_excerpt":"In this paper, we present an approach for evaluating attribution maps, which play a central role in interpreting the predictions of convolutional neural networks (CNNs). We show that the widely used insertion/deletion metrics are susceptible to distribution shifts that affect the reliability of the ranking. Our method proposes to replace pixel modifications with adversarial perturbations, which provides a more robust evaluation framework. By using smoothness and monotonicity measures, we illustrate the effectiveness of our approach in correcting distribution shifts. In addition, we conduct the"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2411.14946","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/2411.14946/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":"2411.14946","created_at":"2026-07-05T09:39:14.037648+00:00"},{"alias_kind":"arxiv_version","alias_value":"2411.14946v1","created_at":"2026-07-05T09:39:14.037648+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2411.14946","created_at":"2026-07-05T09:39:14.037648+00:00"},{"alias_kind":"pith_short_12","alias_value":"RONP62W2ZYUE","created_at":"2026-07-05T09:39:14.037648+00:00"},{"alias_kind":"pith_short_16","alias_value":"RONP62W2ZYUEJMOL","created_at":"2026-07-05T09:39:14.037648+00:00"},{"alias_kind":"pith_short_8","alias_value":"RONP62W2","created_at":"2026-07-05T09:39:14.037648+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/RONP62W2ZYUEJMOL7PNAKGDGMZ","json":"https://pith.science/pith/RONP62W2ZYUEJMOL7PNAKGDGMZ.json","graph_json":"https://pith.science/api/pith-number/RONP62W2ZYUEJMOL7PNAKGDGMZ/graph.json","events_json":"https://pith.science/api/pith-number/RONP62W2ZYUEJMOL7PNAKGDGMZ/events.json","paper":"https://pith.science/paper/RONP62W2"},"agent_actions":{"view_html":"https://pith.science/pith/RONP62W2ZYUEJMOL7PNAKGDGMZ","download_json":"https://pith.science/pith/RONP62W2ZYUEJMOL7PNAKGDGMZ.json","view_paper":"https://pith.science/paper/RONP62W2","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2411.14946&json=true","fetch_graph":"https://pith.science/api/pith-number/RONP62W2ZYUEJMOL7PNAKGDGMZ/graph.json","fetch_events":"https://pith.science/api/pith-number/RONP62W2ZYUEJMOL7PNAKGDGMZ/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/RONP62W2ZYUEJMOL7PNAKGDGMZ/action/timestamp_anchor","attest_storage":"https://pith.science/pith/RONP62W2ZYUEJMOL7PNAKGDGMZ/action/storage_attestation","attest_author":"https://pith.science/pith/RONP62W2ZYUEJMOL7PNAKGDGMZ/action/author_attestation","sign_citation":"https://pith.science/pith/RONP62W2ZYUEJMOL7PNAKGDGMZ/action/citation_signature","submit_replication":"https://pith.science/pith/RONP62W2ZYUEJMOL7PNAKGDGMZ/action/replication_record"}},"created_at":"2026-07-05T09:39:14.037648+00:00","updated_at":"2026-07-05T09:39:14.037648+00:00"}