{"id":"W2977026452","doi":"10.48550/arxiv.1909.12673","title":"A Constructive Prediction of the Generalization Error Across Scales","year":2019,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":51,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University","funders":"","keywords":"Generalization; Dependency (UML); Constructive; Computer science; Construct (python library); Scaling; Scale (ratio); Artificial neural network; Generalization error; Artificial intelligence; Range (aeronautics); Algorithm; Machine learning; Mathematics; Process (computing)","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00007917008,0.0001685956,0.0001867537,0.00005215167,0.0001640985,0.00003067253,0.001414837,0.0001671113,0.000002982269],"category_scores_gemma":[0.0000162337,0.0001566774,0.0001405287,0.000647361,0.0002359746,0.0002735508,0.001810989,0.0002884035,0.0000119963],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001257463,"about_ca_system_score_gemma":0.00008892235,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002276078,"about_ca_topic_score_gemma":0.00002683123,"domain_scores_codex":[0.9987922,0.0001001323,0.000175209,0.0006555877,0.0000874282,0.0001893904],"domain_scores_gemma":[0.9980251,0.00006331437,0.0004026312,0.001263014,0.0002000764,0.00004583427],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00000635452,0.0000247719,0.01170262,0.00002585391,0.00002775865,0.000001533683,0.0001705771,0.8258393,0.0002030915,0.1615572,0.0001258721,0.0003150326],"study_design_scores_gemma":[0.000274876,0.00001741829,0.01999717,0.00006458324,0.00002814029,0.000004632335,0.00005849153,0.9229091,0.0009774464,0.0552045,0.0002830559,0.0001805865],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3343404,0.00001964115,0.6641442,0.00006823239,0.0006018491,0.000382942,0.00007590718,0.00008529314,0.0002815723],"genre_scores_gemma":[0.9970557,0.00005224372,0.002333827,0.00004391423,0.00005731181,0.000002205959,0.00001501274,0.000009875078,0.0004299275],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6627153,"threshold_uncertainty_score":0.638912,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06250519542778074,"score_gpt":0.2126070471350728,"score_spread":0.150101851707292,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}