{"id":"W2996067004","doi":"","title":"Generalization of Two-layer Neural Networks: An Asymptotic Viewpoint","year":2020,"lang":"en","type":"article","venue":"International Conference on Learning Representations","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":34,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Initialization; Generalization; Gradient descent; Layer (electronics); Artificial neural network; Population; Mathematics; Computer science; Flow (mathematics); Applied mathematics; Mathematical analysis; Artificial intelligence; Geometry; Chemistry","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":[0.0001312739,0.0001351424,0.0001470909,0.0001649575,0.0000999498,0.0001755927,0.0007951525,0.00004185088,0.000217608],"category_scores_gemma":[0.0003583629,0.0001413284,0.00006304714,0.0003961961,0.00005498263,0.0006153673,0.0001458241,0.0002075101,0.00001638034],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000325418,"about_ca_system_score_gemma":0.00004535435,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003468486,"about_ca_topic_score_gemma":0.000002480527,"domain_scores_codex":[0.9984488,0.0001663813,0.0003780089,0.0004240898,0.0004351533,0.0001475885],"domain_scores_gemma":[0.998717,0.0001016604,0.0002732697,0.0002738909,0.0005111152,0.0001230418],"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":[0.000008108605,0.00004148878,0.001728785,0.000002200372,0.00001721742,0.000002271201,0.0009353567,0.5710766,0.00116436,0.4224079,0.00008383945,0.002531901],"study_design_scores_gemma":[0.0002881578,0.0002020057,0.001287945,0.00001896936,0.000006782557,0.000004952465,0.00007738733,0.9932674,0.0009568249,0.003717728,0.0000443369,0.0001275149],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.007542129,0.000008076632,0.9818614,0.005441787,0.0002940619,0.000184045,0.000002970623,0.0003131048,0.004352426],"genre_scores_gemma":[0.9681268,0.00001566452,0.03074509,0.0007685212,0.0001164217,0.00003347968,0.00007537912,0.00001318772,0.0001055089],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9605846,"threshold_uncertainty_score":0.5763205,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08250555733982355,"score_gpt":0.3510375405511749,"score_spread":0.2685319832113514,"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."}}