{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01747148,0.001925975,0.001580787,0.002306973,0.0009656179,0.002186561,0.002552228,0.00284974,0.003403918],"category_scores_gemma":[0.1190558,0.001174889,0.00138848,0.001311241,0.005509052,0.007958414,0.004969679,0.005378149,0.0006994525],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001870917,"about_ca_system_score_gemma":0.00152249,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002528592,"about_ca_topic_score_gemma":0.001945393,"domain_scores_codex":[0.9949932,0.002291005,0.0002298965,0.001307385,0.0009184597,0.0002601566],"domain_scores_gemma":[0.919215,0.06189935,0.004433944,0.008299046,0.004933067,0.001219731],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001731026,0.0001328,0.008747489,0.0005904313,0.0002150491,0.0004660493,0.0008442356,0.4895551,0.008308109,0.4134636,0.007818975,0.06968503],"study_design_scores_gemma":[0.00001370005,0.00005363554,0.001632178,0.00008379724,0.00002133849,0.0001229045,0.00005563381,0.8280914,0.001163666,0.1677871,0.0009459264,0.00002859564],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03351824,0.0007116294,0.9577771,0.002679497,0.00012283,0.00006090724,0.000184716,0.0006228936,0.004322059],"genre_scores_gemma":[0.7487125,0.00149291,0.2402678,0.001792511,0.0005523853,0.0005082519,0.0005318016,0.001017082,0.00512478],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01747148,"threshold_uncertainty_score":0.09239912,"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."}}