{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2019:UB6K5K2MUY3EJL2XIPECBWYN4L","short_pith_number":"pith:UB6K5K2M","schema_version":"1.0","canonical_sha256":"a07caeab4ca63644af5743c820db0de2d66db0778cf54c904b39327ed1d2b0b3","source":{"kind":"arxiv","id":"1905.03151","version":2},"attestation_state":"computed","paper":{"title":"Unrestricted Permutation forces Extrapolation: Variable Importance Requires at least One More Model, or There Is No Free Variable Importance","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"stat.ME","authors_text":"Giles Hooker, Lucas Mentch, Siyu Zhou","submitted_at":"2019-05-01T22:37:47Z","abstract_excerpt":"This paper reviews and advocates against the use of permute-and-predict (PaP) methods for interpreting black box functions. Methods such as the variable importance measures proposed for random forests, partial dependence plots, and individual conditional expectation plots remain popular because they are both model-agnostic and depend only on the pre-trained model output, making them computationally efficient and widely available in software. However, numerous studies have found that these tools can produce diagnostics that are highly misleading, particularly when there is strong dependence amo"},"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":"1905.03151","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"stat.ME","submitted_at":"2019-05-01T22:37:47Z","cross_cats_sorted":["cs.LG","stat.ML"],"title_canon_sha256":"c93d80a9cab8775a85ec3469bc56265f11925968b37271953a9383df246d8436","abstract_canon_sha256":"b8aed8a8a0ffeec63ed28e17abfff98dd1a84d37fb0aabf1ded86b767241550a"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:20:44.857373Z","signature_b64":"IxHrHDNcVrCG2N50sJVGla5ieFAKx+g7sgiPvK6ZZ+pISgNWJvNsYCoCvzW2i56qXKvunJWjPkUR0TxzyRQcDA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a07caeab4ca63644af5743c820db0de2d66db0778cf54c904b39327ed1d2b0b3","last_reissued_at":"2026-07-05T03:20:44.856905Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:20:44.856905Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unrestricted Permutation forces Extrapolation: Variable Importance Requires at least One More Model, or There Is No Free Variable Importance","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.LG","stat.ML"],"primary_cat":"stat.ME","authors_text":"Giles Hooker, Lucas Mentch, Siyu Zhou","submitted_at":"2019-05-01T22:37:47Z","abstract_excerpt":"This paper reviews and advocates against the use of permute-and-predict (PaP) methods for interpreting black box functions. Methods such as the variable importance measures proposed for random forests, partial dependence plots, and individual conditional expectation plots remain popular because they are both model-agnostic and depend only on the pre-trained model output, making them computationally efficient and widely available in software. However, numerous studies have found that these tools can produce diagnostics that are highly misleading, particularly when there is strong dependence amo"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1905.03151","kind":"arxiv","version":2},"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/1905.03151/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":"1905.03151","created_at":"2026-07-05T03:20:44.856954+00:00"},{"alias_kind":"arxiv_version","alias_value":"1905.03151v2","created_at":"2026-07-05T03:20:44.856954+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1905.03151","created_at":"2026-07-05T03:20:44.856954+00:00"},{"alias_kind":"pith_short_12","alias_value":"UB6K5K2MUY3E","created_at":"2026-07-05T03:20:44.856954+00:00"},{"alias_kind":"pith_short_16","alias_value":"UB6K5K2MUY3EJL2X","created_at":"2026-07-05T03:20:44.856954+00:00"},{"alias_kind":"pith_short_8","alias_value":"UB6K5K2M","created_at":"2026-07-05T03:20:44.856954+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2509.00846","citing_title":"Causal SHAP: Feature Attribution with Dependency Awareness through Causal Discovery","ref_index":26,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/UB6K5K2MUY3EJL2XIPECBWYN4L","json":"https://pith.science/pith/UB6K5K2MUY3EJL2XIPECBWYN4L.json","graph_json":"https://pith.science/api/pith-number/UB6K5K2MUY3EJL2XIPECBWYN4L/graph.json","events_json":"https://pith.science/api/pith-number/UB6K5K2MUY3EJL2XIPECBWYN4L/events.json","paper":"https://pith.science/paper/UB6K5K2M"},"agent_actions":{"view_html":"https://pith.science/pith/UB6K5K2MUY3EJL2XIPECBWYN4L","download_json":"https://pith.science/pith/UB6K5K2MUY3EJL2XIPECBWYN4L.json","view_paper":"https://pith.science/paper/UB6K5K2M","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1905.03151&json=true","fetch_graph":"https://pith.science/api/pith-number/UB6K5K2MUY3EJL2XIPECBWYN4L/graph.json","fetch_events":"https://pith.science/api/pith-number/UB6K5K2MUY3EJL2XIPECBWYN4L/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/UB6K5K2MUY3EJL2XIPECBWYN4L/action/timestamp_anchor","attest_storage":"https://pith.science/pith/UB6K5K2MUY3EJL2XIPECBWYN4L/action/storage_attestation","attest_author":"https://pith.science/pith/UB6K5K2MUY3EJL2XIPECBWYN4L/action/author_attestation","sign_citation":"https://pith.science/pith/UB6K5K2MUY3EJL2XIPECBWYN4L/action/citation_signature","submit_replication":"https://pith.science/pith/UB6K5K2MUY3EJL2XIPECBWYN4L/action/replication_record"}},"created_at":"2026-07-05T03:20:44.856954+00:00","updated_at":"2026-07-05T03:20:44.856954+00:00"}