{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:VB6XASACSKJGFO4ZI3FWZWVGYR","short_pith_number":"pith:VB6XASAC","schema_version":"1.0","canonical_sha256":"a87d704802929262bb9946cb6cdaa6c442cc5d43c66e851dae06d861fc6247ed","source":{"kind":"arxiv","id":"2605.15920","version":1},"attestation_state":"computed","paper":{"title":"Unsupervised Domain Shift Detection with Interpretable Subspace Attribution","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"Domain shifts appear as localized density anomalies that can be attributed to small sets of features without using labels.","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Alessandro Laio, Sebastian Springer","submitted_at":"2026-05-15T12:58:00Z","abstract_excerpt":"We developed a tool for detecting domain shifts, namely subtle differences in the probability distributions of datasets. We identify these shifts using an algorithm designed to detect localised density anomalies in high-dimensional feature spaces. If an anomaly is present, we then identify the feature subspace in which the anomaly is most pronounced. This allows us to trace the domain shift to a small set of features, making the shift interpretable. Moreover, we provide a protocol for compensating domain shifts by extracting, from two unlabelled datasets, subsets of samples with no detectable "},"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":true,"formal_links_present":false},"canonical_record":{"source":{"id":"2605.15920","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","primary_cat":"stat.ML","submitted_at":"2026-05-15T12:58:00Z","cross_cats_sorted":["cs.LG"],"title_canon_sha256":"dbeb338d31ed416c305814d11aeeaddb2d9a3d6d187c4c537766e8c042a5df0a","abstract_canon_sha256":"8a8e3aad7051cf512c3f7ef9aef79f96dd25224798872401f446db8278fc5b07"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-20T00:01:45.226290Z","signature_b64":"4BUkx4CdS5h2TrTWn06OzONG/OdhjeXC5qWc722pjmXIF4iULE17yyf4vsnueQhC6VpK8WvxWicmwFT3L1Z8BQ==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a87d704802929262bb9946cb6cdaa6c442cc5d43c66e851dae06d861fc6247ed","last_reissued_at":"2026-05-20T00:01:45.225629Z","signature_status":"signed_v1","first_computed_at":"2026-05-20T00:01:45.225629Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Unsupervised Domain Shift Detection with Interpretable Subspace Attribution","license":"http://creativecommons.org/licenses/by-nc-nd/4.0/","headline":"Domain shifts appear as localized density anomalies that can be attributed to small sets of features without using labels.","cross_cats":["cs.LG"],"primary_cat":"stat.ML","authors_text":"Alessandro Laio, Sebastian Springer","submitted_at":"2026-05-15T12:58:00Z","abstract_excerpt":"We developed a tool for detecting domain shifts, namely subtle differences in the probability distributions of datasets. We identify these shifts using an algorithm designed to detect localised density anomalies in high-dimensional feature spaces. If an anomaly is present, we then identify the feature subspace in which the anomaly is most pronounced. This allows us to trace the domain shift to a small set of features, making the shift interpretable. Moreover, we provide a protocol for compensating domain shifts by extracting, from two unlabelled datasets, subsets of samples with no detectable "},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"The framework recovers both broad and localized shifts together with their supporting feature subspaces on controlled 20-dimensional benchmarks and, when applied to ECG recordings differing in measurement-device composition, detects device-induced shifts and identifies associated ECG features.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The method assumes that domain shifts manifest as detectable localized density anomalies in the chosen feature representation; if the shift is diffuse or only visible after nonlinear transformations not captured by the subspace search, the detection and attribution steps may miss it.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"An unsupervised method detects domain shifts via localized density anomaly search in feature space, attributes the shift to a minimal subspace, and extracts balanced subsets from two unlabeled datasets.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Domain shifts appear as localized density anomalies that can be attributed to small sets of features without using labels.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"ff8168f6d595da9945f059094ce9646a3669743f8173da65ece976e5edbe1fc9"},"source":{"id":"2605.15920","kind":"arxiv","version":1},"verdict":{"id":"5d03e2fc-215e-4cae-983d-724e554c2fad","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T19:22:20.108723Z","strongest_claim":"The framework recovers both broad and localized shifts together with their supporting feature subspaces on controlled 20-dimensional benchmarks and, when applied to ECG recordings differing in measurement-device composition, detects device-induced shifts and identifies associated ECG features.","one_line_summary":"An unsupervised method detects domain shifts via localized density anomaly search in feature space, attributes the shift to a minimal subspace, and extracts balanced subsets from two unlabeled datasets.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The method assumes that domain shifts manifest as detectable localized density anomalies in the chosen feature representation; if the shift is diffuse or only visible after nonlinear transformations not captured by the subspace search, the detection and attribution steps may miss it.","pith_extraction_headline":"Domain shifts appear as localized density anomalies that can be attributed to small sets of features without using labels."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.15920/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T19:31:19.058815Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T19:31:08.516148Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T17:33:46.552547Z","status":"skipped","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T17:01:55.752438Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"ba89d7b20ac881a20889b0927c4ea2b68c4f8db93f8cec5f60d8c7ec0cca8918"},"references":{"count":16,"sample":[{"doi":"","year":null,"title":"Joaquin Quiñonero-Candela, Masashi Sugiyama, Anton Schwaighofer, and Neil D. Lawrence, editors. Dataset Shift in Machine Learning. Neural Information Processing Series. MIT Press, Cambridge, MA,","work_id":"89162bda-c435-4d36-9e9c-748c8b861021","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.7551/mitpress/9780262017091.001.0001","year":2012,"title":"Vision: A Computational Investigation into the Human Representation and Processing of Visual Information","work_id":"f90bc3f0-7c2b-44f6-a8aa-1420c3a84682","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2019,"title":"Stephan Rabanser, Stephan Günnemann, and Zachary C. Lipton. Failing loudly: An empirical study of methods for detecting dataset shift. InAdvances in Neural Information Processing Systems, vol- ume 32,","work_id":"7774d5c8-4c01-4041-9aa2-fb3357327dbc","ref_index":3,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.7551/mitpress/9780262170055.003.0008","year":2009,"title":"Vision: A Computational Investigation into the Human Representation and Processing of Visual Information","work_id":"f90bc3f0-7c2b-44f6-a8aa-1420c3a84682","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"10.1007/s10463-008-0197-x","year":2008,"title":"Direct importance estimation for covariate shift adaptation.Annals of the Institute of Statistical Mathematics, 60(4):699–746, December 2008","work_id":"c9de18d1-25b1-407c-a6e2-7ae55bef5f84","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":16,"snapshot_sha256":"ec24505feeb4f092d627728c570d9d5418cc2fc72ba63483c73204c7a55c40a7","internal_anchors":1},"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":"2605.15920","created_at":"2026-05-20T00:01:45.225723+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.15920v1","created_at":"2026-05-20T00:01:45.225723+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15920","created_at":"2026-05-20T00:01:45.225723+00:00"},{"alias_kind":"pith_short_12","alias_value":"VB6XASACSKJG","created_at":"2026-05-20T00:01:45.225723+00:00"},{"alias_kind":"pith_short_16","alias_value":"VB6XASACSKJGFO4Z","created_at":"2026-05-20T00:01:45.225723+00:00"},{"alias_kind":"pith_short_8","alias_value":"VB6XASAC","created_at":"2026-05-20T00:01:45.225723+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":0,"internal_anchor_count":0,"sample":[]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/VB6XASACSKJGFO4ZI3FWZWVGYR","json":"https://pith.science/pith/VB6XASACSKJGFO4ZI3FWZWVGYR.json","graph_json":"https://pith.science/api/pith-number/VB6XASACSKJGFO4ZI3FWZWVGYR/graph.json","events_json":"https://pith.science/api/pith-number/VB6XASACSKJGFO4ZI3FWZWVGYR/events.json","paper":"https://pith.science/paper/VB6XASAC"},"agent_actions":{"view_html":"https://pith.science/pith/VB6XASACSKJGFO4ZI3FWZWVGYR","download_json":"https://pith.science/pith/VB6XASACSKJGFO4ZI3FWZWVGYR.json","view_paper":"https://pith.science/paper/VB6XASAC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.15920&json=true","fetch_graph":"https://pith.science/api/pith-number/VB6XASACSKJGFO4ZI3FWZWVGYR/graph.json","fetch_events":"https://pith.science/api/pith-number/VB6XASACSKJGFO4ZI3FWZWVGYR/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/VB6XASACSKJGFO4ZI3FWZWVGYR/action/timestamp_anchor","attest_storage":"https://pith.science/pith/VB6XASACSKJGFO4ZI3FWZWVGYR/action/storage_attestation","attest_author":"https://pith.science/pith/VB6XASACSKJGFO4ZI3FWZWVGYR/action/author_attestation","sign_citation":"https://pith.science/pith/VB6XASACSKJGFO4ZI3FWZWVGYR/action/citation_signature","submit_replication":"https://pith.science/pith/VB6XASACSKJGFO4ZI3FWZWVGYR/action/replication_record"}},"created_at":"2026-05-20T00:01:45.225723+00:00","updated_at":"2026-05-20T00:01:45.225723+00:00"}