{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2021:URODOY7JG2W5W35CGZZSTAUS7S","short_pith_number":"pith:URODOY7J","schema_version":"1.0","canonical_sha256":"a45c3763e936addb6fa23673298292fcadb9436c083ffe6446f12625f191e399","source":{"kind":"arxiv","id":"2109.05257","version":2},"attestation_state":"computed","paper":{"title":"Towards a Rigorous Evaluation of Time-series Anomaly Detection","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","stat.ME"],"primary_cat":"cs.LG","authors_text":"Byunghan Lee, Hyun-Soo Choi, Kukjin Choi, Siwon Kim, Sungroh Yoon","submitted_at":"2021-09-11T11:14:04Z","abstract_excerpt":"In recent years, proposed studies on time-series anomaly detection (TAD) report high F1 scores on benchmark TAD datasets, giving the impression of clear improvements in TAD. However, most studies apply a peculiar evaluation protocol called point adjustment (PA) before scoring. In this paper, we theoretically and experimentally reveal that the PA protocol has a great possibility of overestimating the detection performance; that is, even a random anomaly score can easily turn into a state-of-the-art TAD method. Therefore, the comparison of TAD methods after applying the PA protocol can lead to m"},"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":"2109.05257","kind":"arxiv","version":2},"metadata":{"license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","primary_cat":"cs.LG","submitted_at":"2021-09-11T11:14:04Z","cross_cats_sorted":["cs.AI","stat.ME"],"title_canon_sha256":"5dd56ef416a1d37584538457ee0664a5b23586f85aa57eb8465ddcd8de3c2925","abstract_canon_sha256":"49dd12d2efcfa671bc9115b3356cadb80c8ec018cb6b343b0074044fc4b72d39"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T03:45:11.789680Z","signature_b64":"tdg6y3hwkR/XuHavYIzRuvj+WKdWWWFpCCQGu67ehXNN6agAx6UTHupV+gkQQ1ejoreQTW14b1hRRu61IM7WCA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"a45c3763e936addb6fa23673298292fcadb9436c083ffe6446f12625f191e399","last_reissued_at":"2026-07-05T03:45:11.789126Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T03:45:11.789126Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Towards a Rigorous Evaluation of Time-series Anomaly Detection","license":"http://creativecommons.org/licenses/by-nc-sa/4.0/","headline":"","cross_cats":["cs.AI","stat.ME"],"primary_cat":"cs.LG","authors_text":"Byunghan Lee, Hyun-Soo Choi, Kukjin Choi, Siwon Kim, Sungroh Yoon","submitted_at":"2021-09-11T11:14:04Z","abstract_excerpt":"In recent years, proposed studies on time-series anomaly detection (TAD) report high F1 scores on benchmark TAD datasets, giving the impression of clear improvements in TAD. However, most studies apply a peculiar evaluation protocol called point adjustment (PA) before scoring. In this paper, we theoretically and experimentally reveal that the PA protocol has a great possibility of overestimating the detection performance; that is, even a random anomaly score can easily turn into a state-of-the-art TAD method. Therefore, the comparison of TAD methods after applying the PA protocol can lead to m"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2109.05257","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/2109.05257/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":"2109.05257","created_at":"2026-07-05T03:45:11.789197+00:00"},{"alias_kind":"arxiv_version","alias_value":"2109.05257v2","created_at":"2026-07-05T03:45:11.789197+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2109.05257","created_at":"2026-07-05T03:45:11.789197+00:00"},{"alias_kind":"pith_short_12","alias_value":"URODOY7JG2W5","created_at":"2026-07-05T03:45:11.789197+00:00"},{"alias_kind":"pith_short_16","alias_value":"URODOY7JG2W5W35C","created_at":"2026-07-05T03:45:11.789197+00:00"},{"alias_kind":"pith_short_8","alias_value":"URODOY7J","created_at":"2026-07-05T03:45:11.789197+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":5,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2607.06094","citing_title":"Modeling Normal Is All You Need: Joint Latent Clustering for Anomaly Detection in Multimodal Cyber-Physical Systems","ref_index":1,"is_internal_anchor":true},{"citing_arxiv_id":"2607.00720","citing_title":"Detecting the Undetectable: Enhancing Unsupervised time series Anomaly Detection via Active Learning","ref_index":24,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11130","citing_title":"HEPA: A Self-Supervised Horizon-Conditioned Event Predictive Architecture for Time Series","ref_index":31,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11130","citing_title":"HEPA: A Self-Supervised Horizon-Conditioned Event Predictive Architecture for Time Series","ref_index":29,"is_internal_anchor":false},{"citing_arxiv_id":"2605.11130","citing_title":"HEPA: A Self-Supervised Horizon-Conditioned Event Predictive Architecture for Time Series","ref_index":29,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/URODOY7JG2W5W35CGZZSTAUS7S","json":"https://pith.science/pith/URODOY7JG2W5W35CGZZSTAUS7S.json","graph_json":"https://pith.science/api/pith-number/URODOY7JG2W5W35CGZZSTAUS7S/graph.json","events_json":"https://pith.science/api/pith-number/URODOY7JG2W5W35CGZZSTAUS7S/events.json","paper":"https://pith.science/paper/URODOY7J"},"agent_actions":{"view_html":"https://pith.science/pith/URODOY7JG2W5W35CGZZSTAUS7S","download_json":"https://pith.science/pith/URODOY7JG2W5W35CGZZSTAUS7S.json","view_paper":"https://pith.science/paper/URODOY7J","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2109.05257&json=true","fetch_graph":"https://pith.science/api/pith-number/URODOY7JG2W5W35CGZZSTAUS7S/graph.json","fetch_events":"https://pith.science/api/pith-number/URODOY7JG2W5W35CGZZSTAUS7S/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/URODOY7JG2W5W35CGZZSTAUS7S/action/timestamp_anchor","attest_storage":"https://pith.science/pith/URODOY7JG2W5W35CGZZSTAUS7S/action/storage_attestation","attest_author":"https://pith.science/pith/URODOY7JG2W5W35CGZZSTAUS7S/action/author_attestation","sign_citation":"https://pith.science/pith/URODOY7JG2W5W35CGZZSTAUS7S/action/citation_signature","submit_replication":"https://pith.science/pith/URODOY7JG2W5W35CGZZSTAUS7S/action/replication_record"}},"created_at":"2026-07-05T03:45:11.789197+00:00","updated_at":"2026-07-05T03:45:11.789197+00:00"}