{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2018:W6CBFXAU6UYUXLGT6OBCBUDXC3","short_pith_number":"pith:W6CBFXAU","schema_version":"1.0","canonical_sha256":"b78412dc14f5314bacd3f38220d07716ee1f6f577063fc6db470c6a331dfd654","source":{"kind":"arxiv","id":"1802.03888","version":3},"attestation_state":"computed","paper":{"title":"Consistent Individualized Feature Attribution for Tree Ensembles","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Gabriel G. Erion, Scott M. Lundberg, Su-In Lee","submitted_at":"2018-02-12T04:23:03Z","abstract_excerpt":"Interpreting predictions from tree ensemble methods such as gradient boosting machines and random forests is important, yet feature attribution for trees is often heuristic and not individualized for each prediction. Here we show that popular feature attribution methods are inconsistent, meaning they can lower a feature's assigned importance when the true impact of that feature actually increases. This is a fundamental problem that casts doubt on any comparison between features. To address it we turn to recent applications of game theory and develop fast exact tree solutions for SHAP (SHapley "},"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":"1802.03888","kind":"arxiv","version":3},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2018-02-12T04:23:03Z","cross_cats_sorted":["stat.ML"],"title_canon_sha256":"93374f06aa48e926928e6258cd6619a91d521a6dcee41197e2c6512ca574b3fb","abstract_canon_sha256":"679b65bf1c41e5e92764383e7afeb7f0daabfa7813a38a03e48a6706b5dba594"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:51:54.768540Z","signature_b64":"H46yUdKwaXDqnxImOWpKQmgd7JzGJ6pCOL0KMOIPKwg1JpNCWCgrQ8vy+EB2hQif+TWUF9N8EHdkAzIIHqUFDQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"b78412dc14f5314bacd3f38220d07716ee1f6f577063fc6db470c6a331dfd654","last_reissued_at":"2026-05-17T23:51:54.768100Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:51:54.768100Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Consistent Individualized Feature Attribution for Tree Ensembles","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["stat.ML"],"primary_cat":"cs.LG","authors_text":"Gabriel G. Erion, Scott M. Lundberg, Su-In Lee","submitted_at":"2018-02-12T04:23:03Z","abstract_excerpt":"Interpreting predictions from tree ensemble methods such as gradient boosting machines and random forests is important, yet feature attribution for trees is often heuristic and not individualized for each prediction. Here we show that popular feature attribution methods are inconsistent, meaning they can lower a feature's assigned importance when the true impact of that feature actually increases. This is a fundamental problem that casts doubt on any comparison between features. To address it we turn to recent applications of game theory and develop fast exact tree solutions for SHAP (SHapley "},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"1802.03888","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":""},"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":"1802.03888","created_at":"2026-05-17T23:51:54.768167+00:00"},{"alias_kind":"arxiv_version","alias_value":"1802.03888v3","created_at":"2026-05-17T23:51:54.768167+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.1802.03888","created_at":"2026-05-17T23:51:54.768167+00:00"},{"alias_kind":"pith_short_12","alias_value":"W6CBFXAU6UYU","created_at":"2026-05-18T12:32:59.047623+00:00"},{"alias_kind":"pith_short_16","alias_value":"W6CBFXAU6UYUXLGT","created_at":"2026-05-18T12:32:59.047623+00:00"},{"alias_kind":"pith_short_8","alias_value":"W6CBFXAU","created_at":"2026-05-18T12:32:59.047623+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":14,"internal_anchor_count":7,"sample":[{"citing_arxiv_id":"1907.03324","citing_title":"A Human-Grounded Evaluation of SHAP for Alert Processing","ref_index":12,"is_internal_anchor":true},{"citing_arxiv_id":"1907.03334","citing_title":"Case-Based Reasoning for Assisting Domain Experts in Processing Fraud Alerts of Black-Box Machine Learning Models","ref_index":9,"is_internal_anchor":true},{"citing_arxiv_id":"2605.18422","citing_title":"Generalized Functional ANOVA in Closed-Form: A Unified View of Additive Explanations","ref_index":35,"is_internal_anchor":true},{"citing_arxiv_id":"2506.09255","citing_title":"AI-Driven SEEG Channel Ranking for Epileptogenic Zone Localization","ref_index":17,"is_internal_anchor":true},{"citing_arxiv_id":"2511.09376","citing_title":"From Decision Trees to Boolean Logic: A Fast and Unified SHAP Algorithm","ref_index":2,"is_internal_anchor":true},{"citing_arxiv_id":"2603.06011","citing_title":"Spectra-Scope : A toolkit for automated and interpretable characterization of material properties from spectral data","ref_index":24,"is_internal_anchor":true},{"citing_arxiv_id":"2603.26723","citing_title":"Interpretable liquid crystal phase classification via two-by-two ordinal patterns","ref_index":57,"is_internal_anchor":true},{"citing_arxiv_id":"2605.12281","citing_title":"What makes a word hard to learn? Modeling L1 influence on English vocabulary difficulty","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2605.10598","citing_title":"Budget-Efficient Automatic Algorithm Design via Code Graph","ref_index":26,"is_internal_anchor":false},{"citing_arxiv_id":"2604.23500","citing_title":"Interpretable Physics-Informed Load Forecasting for U.S. Grid Resilience: SHAP-Guided Ensemble Validation in Hybrid Deep Learning Under Extreme Weather","ref_index":16,"is_internal_anchor":false},{"citing_arxiv_id":"2604.22662","citing_title":"Rethinking XAI Evaluation: A Human-Centered Audit of Shapley Benchmarks in High-Stakes Settings","ref_index":35,"is_internal_anchor":false},{"citing_arxiv_id":"2605.01189","citing_title":"NEURON: A Neuro-symbolic System for Grounded Clinical Explainability","ref_index":62,"is_internal_anchor":false},{"citing_arxiv_id":"2604.05857","citing_title":"Weight-Informed Self-Explaining Clustering for Mixed-Type Tabular Data","ref_index":4,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13462","citing_title":"Learning from Change: Predictive Models for Incident Prevention in a Regulated IT Environment","ref_index":28,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/W6CBFXAU6UYUXLGT6OBCBUDXC3","json":"https://pith.science/pith/W6CBFXAU6UYUXLGT6OBCBUDXC3.json","graph_json":"https://pith.science/api/pith-number/W6CBFXAU6UYUXLGT6OBCBUDXC3/graph.json","events_json":"https://pith.science/api/pith-number/W6CBFXAU6UYUXLGT6OBCBUDXC3/events.json","paper":"https://pith.science/paper/W6CBFXAU"},"agent_actions":{"view_html":"https://pith.science/pith/W6CBFXAU6UYUXLGT6OBCBUDXC3","download_json":"https://pith.science/pith/W6CBFXAU6UYUXLGT6OBCBUDXC3.json","view_paper":"https://pith.science/paper/W6CBFXAU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=1802.03888&json=true","fetch_graph":"https://pith.science/api/pith-number/W6CBFXAU6UYUXLGT6OBCBUDXC3/graph.json","fetch_events":"https://pith.science/api/pith-number/W6CBFXAU6UYUXLGT6OBCBUDXC3/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/W6CBFXAU6UYUXLGT6OBCBUDXC3/action/timestamp_anchor","attest_storage":"https://pith.science/pith/W6CBFXAU6UYUXLGT6OBCBUDXC3/action/storage_attestation","attest_author":"https://pith.science/pith/W6CBFXAU6UYUXLGT6OBCBUDXC3/action/author_attestation","sign_citation":"https://pith.science/pith/W6CBFXAU6UYUXLGT6OBCBUDXC3/action/citation_signature","submit_replication":"https://pith.science/pith/W6CBFXAU6UYUXLGT6OBCBUDXC3/action/replication_record"}},"created_at":"2026-05-17T23:51:54.768167+00:00","updated_at":"2026-05-17T23:51:54.768167+00:00"}