{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:NGJCUYVMTXYIAEUQPIFXOXPXXV","short_pith_number":"pith:NGJCUYVM","schema_version":"1.0","canonical_sha256":"69922a62ac9df08012907a0b775df7bd553899de932984daf5eb45a7b0f14d6c","source":{"kind":"arxiv","id":"1806.10758","version":3},"attestation_state":"computed","paper":{"title":"A Benchmark for Interpretability Methods in Deep Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Been Kim, Dumitru Erhan, Pieter-Jan Kindermans, Sara Hooker","submitted_at":"2018-06-28T03:46:57Z","abstract_excerpt":"We propose an empirical measure of the approximate accuracy of feature importance estimates in deep neural networks. Our results across several large-scale image classification datasets show that many popular interpretability methods produce estimates of feature importance that are not better than a random designation of feature importance. Only certain ensemble based approaches---VarGrad and SmoothGrad-Squared---outperform such a random assignment of importance. The manner of ensembling remains critical, we show that some approaches do no better then the underlying method but carry a far high"},"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":"1806.10758","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-06-28T03:46:57Z","cross_cats_sorted":["cs.AI","stat.ML"],"title_canon_sha256":"9ee7f61f4b76e8b3baada9a2a7f0e937012a7d666140b8259352d9512ed32352","abstract_canon_sha256":"15ceae11b3fd2fa2c64bfa18fa522268536178f0bf196717f545d66fa59d1224"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T00:16:58.044103Z","signature_b64":"BFegxul2hkmDnxonLAVqIkWrC4HKQ+06QYpKrH5ckJJeMaKxL8HEtn+Nh9AmTblcmlqfdnUpiHIhilZr3lOzCg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"69922a62ac9df08012907a0b775df7bd553899de932984daf5eb45a7b0f14d6c","last_reissued_at":"2026-07-05T00:16:58.043642Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T00:16:58.043642Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"A Benchmark for Interpretability Methods in Deep Neural Networks","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.AI","stat.ML"],"primary_cat":"cs.LG","authors_text":"Been Kim, Dumitru Erhan, Pieter-Jan Kindermans, Sara Hooker","submitted_at":"2018-06-28T03:46:57Z","abstract_excerpt":"We propose an empirical measure of the approximate accuracy of feature importance estimates in deep neural networks. Our results across several large-scale image classification datasets show that many popular interpretability methods produce estimates of feature importance that are not better than a random designation of feature importance. Only certain ensemble based approaches---VarGrad and SmoothGrad-Squared---outperform such a random assignment of importance. The manner of ensembling remains critical, we show that some approaches do no better then the underlying method but carry a far high"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1806.10758","kind":"arxiv","version":3},"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/1806.10758/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":"1806.10758","created_at":"2026-07-05T00:16:58.043699+00:00"},{"alias_kind":"arxiv_version","alias_value":"1806.10758v3","created_at":"2026-07-05T00:16:58.043699+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1806.10758","created_at":"2026-07-05T00:16:58.043699+00:00"},{"alias_kind":"pith_short_12","alias_value":"NGJCUYVMTXYI","created_at":"2026-07-05T00:16:58.043699+00:00"},{"alias_kind":"pith_short_16","alias_value":"NGJCUYVMTXYIAEUQ","created_at":"2026-07-05T00:16:58.043699+00:00"},{"alias_kind":"pith_short_8","alias_value":"NGJCUYVM","created_at":"2026-07-05T00:16:58.043699+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.28391","citing_title":"Few-class Fidelity: Evaluating Explanations of Real-conditions CNN classifiers with Optimized Perturbations","ref_index":12,"is_internal_anchor":false},{"citing_arxiv_id":"2605.28977","citing_title":"Comparing Post-Hoc Explainable AI Methods for Interpreting Black-Box EEG Models in Depression Detection","ref_index":67,"is_internal_anchor":false},{"citing_arxiv_id":"2605.16905","citing_title":"AIM: Adversarial Information Masking for Faithfulness Evaluation of Saliency Maps","ref_index":23,"is_internal_anchor":false},{"citing_arxiv_id":"2509.09794","citing_title":"Synthetic Homes: A Multimodal Generative AI Pipeline for Residential Building Data Generation under Data Scarcity","ref_index":43,"is_internal_anchor":false},{"citing_arxiv_id":"2604.08502","citing_title":"Quantifying Explanation Consistency: The C-Score Metric for CAM-Based Explainability in Medical Image Classification","ref_index":35,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/NGJCUYVMTXYIAEUQPIFXOXPXXV","json":"https://pith.science/pith/NGJCUYVMTXYIAEUQPIFXOXPXXV.json","graph_json":"https://pith.science/api/pith-number/NGJCUYVMTXYIAEUQPIFXOXPXXV/graph.json","events_json":"https://pith.science/api/pith-number/NGJCUYVMTXYIAEUQPIFXOXPXXV/events.json","paper":"https://pith.science/paper/NGJCUYVM"},"agent_actions":{"view_html":"https://pith.science/pith/NGJCUYVMTXYIAEUQPIFXOXPXXV","download_json":"https://pith.science/pith/NGJCUYVMTXYIAEUQPIFXOXPXXV.json","view_paper":"https://pith.science/paper/NGJCUYVM","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1806.10758&json=true","fetch_graph":"https://pith.science/api/pith-number/NGJCUYVMTXYIAEUQPIFXOXPXXV/graph.json","fetch_events":"https://pith.science/api/pith-number/NGJCUYVMTXYIAEUQPIFXOXPXXV/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/NGJCUYVMTXYIAEUQPIFXOXPXXV/action/timestamp_anchor","attest_storage":"https://pith.science/pith/NGJCUYVMTXYIAEUQPIFXOXPXXV/action/storage_attestation","attest_author":"https://pith.science/pith/NGJCUYVMTXYIAEUQPIFXOXPXXV/action/author_attestation","sign_citation":"https://pith.science/pith/NGJCUYVMTXYIAEUQPIFXOXPXXV/action/citation_signature","submit_replication":"https://pith.science/pith/NGJCUYVMTXYIAEUQPIFXOXPXXV/action/replication_record"}},"created_at":"2026-07-05T00:16:58.043699+00:00","updated_at":"2026-07-05T00:16:58.043699+00:00"}