{"state_type":"pith_open_graph_state","state_version":"1.0","pith_number":"pith:2026:HPZFVXZ5XMNIQA3QEML5XFVCM6","merge_version":"pith-open-graph-merge-v1","event_count":2,"valid_event_count":2,"invalid_event_count":0,"equivocation_count":0,"current":{"canonical_record":{"metadata":{"abstract_canon_sha256":"b891ca503526fd49ecbca3882f38b56826d868c585390c40ded014ea24ade2e1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-10T18:03:19Z","title_canon_sha256":"55bb0b2220e4706bf927ba9288d6fe9843217b4d04fdf223ae52fb238c13ef66"},"schema_version":"1.0","source":{"id":"2606.12552","kind":"arxiv","version":1}},"source_aliases":[{"alias_kind":"arxiv","alias_value":"2606.12552","created_at":"2026-06-12T00:07:54Z"},{"alias_kind":"arxiv_version","alias_value":"2606.12552v1","created_at":"2026-06-12T00:07:54Z"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2606.12552","created_at":"2026-06-12T00:07:54Z"},{"alias_kind":"pith_short_12","alias_value":"HPZFVXZ5XMNI","created_at":"2026-06-12T00:07:54Z"},{"alias_kind":"pith_short_16","alias_value":"HPZFVXZ5XMNIQA3Q","created_at":"2026-06-12T00:07:54Z"},{"alias_kind":"pith_short_8","alias_value":"HPZFVXZ5","created_at":"2026-06-12T00:07:54Z"}],"graph_snapshots":[{"event_id":"sha256:4945ee6680eb52903aee8cfcdd6285949fb95524fb6b2d2bffbb76c87916e633","target":"graph","created_at":"2026-06-12T00:07:54Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"graph_snapshot":{"author_claims":{"count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","strong_count":0},"builder_version":"pith-number-builder-2026-05-17-v1","claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"integrity":{"available":true,"clean":true,"detectors_run":[],"endpoint":"/pith/2606.12552/integrity.json","findings":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938","summary":{"advisory":0,"by_detector":{},"critical":0,"informational":0}},"paper":{"abstract_excerpt":"Modern machine learning progresses through empirical work, benchmarking new methods to evaluate relative performance. However, the statistical variability inherent to evaluation - exacerbated by the stochastic nature of many algorithms - often makes performance estimation unreliable due to the limited test samples available, leading to a validation crisis in which genuine advances are difficult to discern. In this work, we show that cross-validation improves markedly confidence when evaluating and comparing learning algorithm performances. We introduce the concept of sample gain, which quantif","authors_text":"C\\'elestin Eve, Ga\\\"el Varoquaux, Thomas Moreau","cross_cats":[],"headline":"","license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-10T18:03:19Z","title":"Crossing the Validation Crisis: Cross-Validation Reduces Benchmarking Variance Surprisingly Well"},"references":{"count":0,"internal_anchors":0,"resolved_work":0,"sample":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2606.12552","kind":"arxiv","version":1},"verdict":{"created_at":null,"id":null,"model_set":{},"one_line_summary":"","pipeline_version":null,"pith_extraction_headline":"","strongest_claim":"","weakest_assumption":""}},"verdict_id":null}}],"author_attestations":[],"timestamp_anchors":[],"storage_attestations":[],"citation_signatures":[],"replication_records":[],"corrections":[],"mirror_hints":[],"record_created":{"event_id":"sha256:f42b2bcdf576e1454a937d3554b32d02e6dfff4313d936721c138afcccf96add","target":"record","created_at":"2026-06-12T00:07:54Z","signer":{"key_id":"pith-v1-2026-05","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","signer_id":"pith.science","signer_type":"pith_registry"},"payload":{"attestation_state":"computed","canonical_record":{"metadata":{"abstract_canon_sha256":"b891ca503526fd49ecbca3882f38b56826d868c585390c40ded014ea24ade2e1","cross_cats_sorted":[],"license":"http://creativecommons.org/licenses/by/4.0/","primary_cat":"cs.LG","submitted_at":"2026-06-10T18:03:19Z","title_canon_sha256":"55bb0b2220e4706bf927ba9288d6fe9843217b4d04fdf223ae52fb238c13ef66"},"schema_version":"1.0","source":{"id":"2606.12552","kind":"arxiv","version":1}},"canonical_sha256":"3bf25adf3dbb1a8803702317db96a267932d9531b34d13b5dd06033aa779b119","receipt":{"algorithm":"ed25519","builder_version":"pith-number-builder-2026-05-17-v1","canonical_sha256":"3bf25adf3dbb1a8803702317db96a267932d9531b34d13b5dd06033aa779b119","first_computed_at":"2026-06-12T00:07:54.375849Z","key_id":"pith-v1-2026-05","kind":"pith_receipt","last_reissued_at":"2026-06-12T00:07:54.375849Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54","receipt_version":"0.3","signature_b64":"whGCTMx6ofIk/XeS1M6d82Tc0szb1gbR8z87ugsEqJoCD94cdEH/OWYO77//B2Gm0Kf69WtTL36JUxp22/I3AA==","signature_status":"signed_v1","signed_at":"2026-06-12T00:07:54.376235Z","signed_message":"canonical_sha256_bytes"},"source_id":"2606.12552","source_kind":"arxiv","source_version":1}}},"equivocations":[],"invalid_events":[],"applied_event_ids":["sha256:f42b2bcdf576e1454a937d3554b32d02e6dfff4313d936721c138afcccf96add","sha256:4945ee6680eb52903aee8cfcdd6285949fb95524fb6b2d2bffbb76c87916e633"],"state_sha256":"80fe3694413362f7d6e1c0567a4e3e466a2ae0ce3f60aeb7614c55018f9c0fa2"}