{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:GBKOJ74K7NLMMJPBXILPIEU4YP","short_pith_number":"pith:GBKOJ74K","schema_version":"1.0","canonical_sha256":"3054e4ff8afb56c625e1ba16f4129cc3fe395d76cf7993158d6c9c84877aba75","source":{"kind":"arxiv","id":"2203.06877","version":1},"attestation_state":"computed","paper":{"title":"Rethinking Stability for Attribution-based Explanations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chirag Agarwal, Eshika Saxena, Himabindu Lakkaraju, Marinka Zitnik, Martin Pawelczyk, Nari Johnson, Satyapriya Krishna","submitted_at":"2022-03-14T06:19:27Z","abstract_excerpt":"As attribution-based explanation methods are increasingly used to establish model trustworthiness in high-stakes situations, it is critical to ensure that these explanations are stable, e.g., robust to infinitesimal perturbations to an input. However, previous works have shown that state-of-the-art explanation methods generate unstable explanations. Here, we introduce metrics to quantify the stability of an explanation and show that several popular explanation methods are unstable. In particular, we propose new Relative Stability metrics that measure the change in output explanation with respe"},"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":"2203.06877","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2022-03-14T06:19:27Z","cross_cats_sorted":[],"title_canon_sha256":"8acae4a78bdee54675f08e99b1bdc9bc60f5436ba8e7acd1e76a95671d33cedf","abstract_canon_sha256":"722c7264f7667c3354737617c65005d1af7a82f2e22275fe27376e878828728d"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:04:57.125838Z","signature_b64":"OXIUlwys7/qjhxWRYW/Wu9RBTTpK5yw7Od28YDY2QIXc4ffyatrH//e7mXITXmqH0/eRz1gzoUJ6uRY29o5qDw==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"3054e4ff8afb56c625e1ba16f4129cc3fe395d76cf7993158d6c9c84877aba75","last_reissued_at":"2026-07-05T04:04:57.125266Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:04:57.125266Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Rethinking Stability for Attribution-based Explanations","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Chirag Agarwal, Eshika Saxena, Himabindu Lakkaraju, Marinka Zitnik, Martin Pawelczyk, Nari Johnson, Satyapriya Krishna","submitted_at":"2022-03-14T06:19:27Z","abstract_excerpt":"As attribution-based explanation methods are increasingly used to establish model trustworthiness in high-stakes situations, it is critical to ensure that these explanations are stable, e.g., robust to infinitesimal perturbations to an input. However, previous works have shown that state-of-the-art explanation methods generate unstable explanations. Here, we introduce metrics to quantify the stability of an explanation and show that several popular explanation methods are unstable. In particular, we propose new Relative Stability metrics that measure the change in output explanation with respe"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2203.06877","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/2203.06877/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":"2203.06877","created_at":"2026-07-05T04:04:57.125339+00:00"},{"alias_kind":"arxiv_version","alias_value":"2203.06877v1","created_at":"2026-07-05T04:04:57.125339+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2203.06877","created_at":"2026-07-05T04:04:57.125339+00:00"},{"alias_kind":"pith_short_12","alias_value":"GBKOJ74K7NLM","created_at":"2026-07-05T04:04:57.125339+00:00"},{"alias_kind":"pith_short_16","alias_value":"GBKOJ74K7NLMMJPB","created_at":"2026-07-05T04:04:57.125339+00:00"},{"alias_kind":"pith_short_8","alias_value":"GBKOJ74K","created_at":"2026-07-05T04:04:57.125339+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2604.22018","citing_title":"Foundation models for discovering robust biomarkers of neurological disorders from dynamic functional connectivity","ref_index":16,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/GBKOJ74K7NLMMJPBXILPIEU4YP","json":"https://pith.science/pith/GBKOJ74K7NLMMJPBXILPIEU4YP.json","graph_json":"https://pith.science/api/pith-number/GBKOJ74K7NLMMJPBXILPIEU4YP/graph.json","events_json":"https://pith.science/api/pith-number/GBKOJ74K7NLMMJPBXILPIEU4YP/events.json","paper":"https://pith.science/paper/GBKOJ74K"},"agent_actions":{"view_html":"https://pith.science/pith/GBKOJ74K7NLMMJPBXILPIEU4YP","download_json":"https://pith.science/pith/GBKOJ74K7NLMMJPBXILPIEU4YP.json","view_paper":"https://pith.science/paper/GBKOJ74K","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2203.06877&json=true","fetch_graph":"https://pith.science/api/pith-number/GBKOJ74K7NLMMJPBXILPIEU4YP/graph.json","fetch_events":"https://pith.science/api/pith-number/GBKOJ74K7NLMMJPBXILPIEU4YP/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/GBKOJ74K7NLMMJPBXILPIEU4YP/action/timestamp_anchor","attest_storage":"https://pith.science/pith/GBKOJ74K7NLMMJPBXILPIEU4YP/action/storage_attestation","attest_author":"https://pith.science/pith/GBKOJ74K7NLMMJPBXILPIEU4YP/action/author_attestation","sign_citation":"https://pith.science/pith/GBKOJ74K7NLMMJPBXILPIEU4YP/action/citation_signature","submit_replication":"https://pith.science/pith/GBKOJ74K7NLMMJPBXILPIEU4YP/action/replication_record"}},"created_at":"2026-07-05T04:04:57.125339+00:00","updated_at":"2026-07-05T04:04:57.125339+00:00"}