{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2026:TZBVH2UCH57UFKYOX24RCL4VK2","short_pith_number":"pith:TZBVH2UC","schema_version":"1.0","canonical_sha256":"9e4353ea823f7f42ab0ebeb9112f9556b566ce5844e266b3081c4e6b79951434","source":{"kind":"arxiv","id":"2605.15134","version":1},"attestation_state":"computed","paper":{"title":"Training ML Models with Predictable Failures","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Scott Niekum, Will Schwarzer","submitted_at":"2026-05-14T17:41:52Z","abstract_excerpt":"Estimating how often an ML model will fail at deployment scale is central to pre-deployment safety assessment, but a feasible evaluation set is rarely large enough to observe the failures that matter. Jones et al. (2025) address this by extrapolating from the largest k failure scores in an evaluation set to predict deployment-scale failure rates. We give a finite-k decomposition of this estimator's forecast error and show that it has a built-in bias toward over-prediction in the typical case, which is the safety-favorable direction. This bias is offset when the evaluation set misses a rare hig"},"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":"2605.15134","kind":"arxiv","version":1},"metadata":{"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-05-14T17:41:52Z","cross_cats_sorted":[],"title_canon_sha256":"e525bd4e7cf044c16fecf61182f97d46ba200441ae04e2b8b1138161726b9511","abstract_canon_sha256":"545b0dff904342ad808e2f41d6a1ec574f3e58016c2667e9567b51e7e3517f6b"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-05-17T23:38:54.021505Z","signature_b64":"EputUrJv/Iuli3IQ41hazMYODsjXSpm5psFG75FIeM65ySMSl4u6O6tIXn/e4F5vNGSp29XEm69su14c8CX0AA==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"9e4353ea823f7f42ab0ebeb9112f9556b566ce5844e266b3081c4e6b79951434","last_reissued_at":"2026-05-17T23:38:54.020847Z","signature_status":"signed_v1","first_computed_at":"2026-05-17T23:38:54.020847Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Training ML Models with Predictable Failures","license":"http://creativecommons.org/licenses/by/4.0/","headline":"","cross_cats":[],"primary_cat":"cs.LG","authors_text":"Scott Niekum, Will Schwarzer","submitted_at":"2026-05-14T17:41:52Z","abstract_excerpt":"Estimating how often an ML model will fail at deployment scale is central to pre-deployment safety assessment, but a feasible evaluation set is rarely large enough to observe the failures that matter. Jones et al. (2025) address this by extrapolating from the largest k failure scores in an evaluation set to predict deployment-scale failure rates. We give a finite-k decomposition of this estimator's forecast error and show that it has a built-in bias toward over-prediction in the typical case, which is the safety-favorable direction. This bias is offset when the evaluation set misses a rare hig"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2605.15134","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":""},"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":"2605.15134","created_at":"2026-05-17T23:38:54.020936+00:00"},{"alias_kind":"arxiv_version","alias_value":"2605.15134v1","created_at":"2026-05-17T23:38:54.020936+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2605.15134","created_at":"2026-05-17T23:38:54.020936+00:00"},{"alias_kind":"pith_short_12","alias_value":"TZBVH2UCH57U","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_16","alias_value":"TZBVH2UCH57UFKYO","created_at":"2026-05-18T12:33:37.589309+00:00"},{"alias_kind":"pith_short_8","alias_value":"TZBVH2UC","created_at":"2026-05-18T12:33:37.589309+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/TZBVH2UCH57UFKYOX24RCL4VK2","json":"https://pith.science/pith/TZBVH2UCH57UFKYOX24RCL4VK2.json","graph_json":"https://pith.science/api/pith-number/TZBVH2UCH57UFKYOX24RCL4VK2/graph.json","events_json":"https://pith.science/api/pith-number/TZBVH2UCH57UFKYOX24RCL4VK2/events.json","paper":"https://pith.science/paper/TZBVH2UC"},"agent_actions":{"view_html":"https://pith.science/pith/TZBVH2UCH57UFKYOX24RCL4VK2","download_json":"https://pith.science/pith/TZBVH2UCH57UFKYOX24RCL4VK2.json","view_paper":"https://pith.science/paper/TZBVH2UC","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2605.15134&json=true","fetch_graph":"https://pith.science/api/pith-number/TZBVH2UCH57UFKYOX24RCL4VK2/graph.json","fetch_events":"https://pith.science/api/pith-number/TZBVH2UCH57UFKYOX24RCL4VK2/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/TZBVH2UCH57UFKYOX24RCL4VK2/action/timestamp_anchor","attest_storage":"https://pith.science/pith/TZBVH2UCH57UFKYOX24RCL4VK2/action/storage_attestation","attest_author":"https://pith.science/pith/TZBVH2UCH57UFKYOX24RCL4VK2/action/author_attestation","sign_citation":"https://pith.science/pith/TZBVH2UCH57UFKYOX24RCL4VK2/action/citation_signature","submit_replication":"https://pith.science/pith/TZBVH2UCH57UFKYOX24RCL4VK2/action/replication_record"}},"created_at":"2026-05-17T23:38:54.020936+00:00","updated_at":"2026-05-17T23:38:54.020936+00:00"}