{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2022:JU33JL6OS3MRG5745FKFPCW56V","short_pith_number":"pith:JU33JL6O","schema_version":"1.0","canonical_sha256":"4d37b4afce96d91377fce954578addf54415b1e453647288d55766bf9cfc8853","source":{"kind":"arxiv","id":"2209.08590","version":1},"attestation_state":"computed","paper":{"title":"RankFeat: Rank-1 Feature Removal for Out-of-distribution Detection","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Nicu Sebe, Wei Wang, Yue Song","submitted_at":"2022-09-18T16:01:31Z","abstract_excerpt":"The task of out-of-distribution (OOD) detection is crucial for deploying machine learning models in real-world settings. In this paper, we observe that the singular value distributions of the in-distribution (ID) and OOD features are quite different: the OOD feature matrix tends to have a larger dominant singular value than the ID feature, and the class predictions of OOD samples are largely determined by it. This observation motivates us to propose \\texttt{RankFeat}, a simple yet effective \\texttt{post hoc} approach for OOD detection by removing the rank-1 matrix composed of the largest singu"},"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":"2209.08590","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2022-09-18T16:01:31Z","cross_cats_sorted":["cs.CV"],"title_canon_sha256":"3e7a5daafce17a9a2cbf61b99273d2e9b4f9cdedf400493c8bb3de3d44c26fc5","abstract_canon_sha256":"7c7a7e09790f0b92676ebbce2bcb5fa4e55648feeae35998df2feaf3db76ce92"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T04:58:43.383197Z","signature_b64":"OELVUNfbpyZ+BjE8wK2Zo54yEdtTN8vmCaPhipQhbfzDAJfzwsV20BeOUlPohR5VM/lhR0SiRMaVx6OqoCZLAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"4d37b4afce96d91377fce954578addf54415b1e453647288d55766bf9cfc8853","last_reissued_at":"2026-07-05T04:58:43.382779Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T04:58:43.382779Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"RankFeat: Rank-1 Feature Removal for Out-of-distribution Detection","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.CV"],"primary_cat":"cs.LG","authors_text":"Nicu Sebe, Wei Wang, Yue Song","submitted_at":"2022-09-18T16:01:31Z","abstract_excerpt":"The task of out-of-distribution (OOD) detection is crucial for deploying machine learning models in real-world settings. In this paper, we observe that the singular value distributions of the in-distribution (ID) and OOD features are quite different: the OOD feature matrix tends to have a larger dominant singular value than the ID feature, and the class predictions of OOD samples are largely determined by it. This observation motivates us to propose \\texttt{RankFeat}, a simple yet effective \\texttt{post hoc} approach for OOD detection by removing the rank-1 matrix composed of the largest singu"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2209.08590","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/2209.08590/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":"2209.08590","created_at":"2026-07-05T04:58:43.382832+00:00"},{"alias_kind":"arxiv_version","alias_value":"2209.08590v1","created_at":"2026-07-05T04:58:43.382832+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2209.08590","created_at":"2026-07-05T04:58:43.382832+00:00"},{"alias_kind":"pith_short_12","alias_value":"JU33JL6OS3MR","created_at":"2026-07-05T04:58:43.382832+00:00"},{"alias_kind":"pith_short_16","alias_value":"JU33JL6OS3MRG574","created_at":"2026-07-05T04:58:43.382832+00:00"},{"alias_kind":"pith_short_8","alias_value":"JU33JL6O","created_at":"2026-07-05T04:58:43.382832+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2506.10089","citing_title":"Optimizing Latent Dimension Allocation in Hierarchical VAEs: Balancing Attenuation and Information Retention for OOD Detection","ref_index":3,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/JU33JL6OS3MRG5745FKFPCW56V","json":"https://pith.science/pith/JU33JL6OS3MRG5745FKFPCW56V.json","graph_json":"https://pith.science/api/pith-number/JU33JL6OS3MRG5745FKFPCW56V/graph.json","events_json":"https://pith.science/api/pith-number/JU33JL6OS3MRG5745FKFPCW56V/events.json","paper":"https://pith.science/paper/JU33JL6O"},"agent_actions":{"view_html":"https://pith.science/pith/JU33JL6OS3MRG5745FKFPCW56V","download_json":"https://pith.science/pith/JU33JL6OS3MRG5745FKFPCW56V.json","view_paper":"https://pith.science/paper/JU33JL6O","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2209.08590&json=true","fetch_graph":"https://pith.science/api/pith-number/JU33JL6OS3MRG5745FKFPCW56V/graph.json","fetch_events":"https://pith.science/api/pith-number/JU33JL6OS3MRG5745FKFPCW56V/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/JU33JL6OS3MRG5745FKFPCW56V/action/timestamp_anchor","attest_storage":"https://pith.science/pith/JU33JL6OS3MRG5745FKFPCW56V/action/storage_attestation","attest_author":"https://pith.science/pith/JU33JL6OS3MRG5745FKFPCW56V/action/author_attestation","sign_citation":"https://pith.science/pith/JU33JL6OS3MRG5745FKFPCW56V/action/citation_signature","submit_replication":"https://pith.science/pith/JU33JL6OS3MRG5745FKFPCW56V/action/replication_record"}},"created_at":"2026-07-05T04:58:43.382832+00:00","updated_at":"2026-07-05T04:58:43.382832+00:00"}