{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.007599839,0.001390821,0.001066081,0.0008918666,0.0005839618,0.001331208,0.002272203,0.001883862,0.002711768],"category_scores_gemma":[0.03536998,0.0007264858,0.001336404,0.0004957762,0.002867395,0.005564086,0.004509371,0.004391959,0.0004302385],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001901266,"about_ca_system_score_gemma":0.000819693,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002248694,"about_ca_topic_score_gemma":0.001353939,"domain_scores_codex":[0.9982773,0.0007747918,0.00007919753,0.0003395945,0.0003699853,0.0001590315],"domain_scores_gemma":[0.9898021,0.006178983,0.0009585062,0.001670659,0.0009851652,0.0004046933],"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":[0.0001375122,0.00006594446,0.002442301,0.0002354233,0.000122262,0.0003075738,0.0002643402,0.5235912,0.003283354,0.4460758,0.002185014,0.02128946],"study_design_scores_gemma":[0.000008639508,0.00004176541,0.0003318472,0.00002841466,0.00001144625,0.00006755155,0.00001832771,0.8780164,0.0004768461,0.1204459,0.000541192,0.00001167852],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04869292,0.0009228216,0.9398533,0.00168622,0.00009381588,0.00005253217,0.0001046587,0.0003927675,0.008201012],"genre_scores_gemma":[0.8704663,0.001534478,0.1150717,0.001296952,0.0004034548,0.0003646893,0.0003707161,0.0003823226,0.01010942],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007599839,"threshold_uncertainty_score":0.04019231,"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."}}