{"id":"W2604117713","doi":"10.48550/arxiv.1703.11008","title":"Computing Nonvacuous Generalization Bounds for Deep (Stochastic) Neural Networks with Many More Parameters than Training Data","year":2017,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":250,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Overfitting; Generalization; Computer science; Artificial neural network; Maxima and minima; Regularization (linguistics); Early stopping; Artificial intelligence; Machine learning; Algorithm; Test data; Mathematics","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0003950736,0.0004992738,0.0005068003,0.000314891,0.0006699547,0.0006544911,0.004854295,0.0003130831,0.000001768902],"category_scores_gemma":[0.0001184184,0.0005474736,0.0001328098,0.0003447445,0.0002556049,0.0007589855,0.003103528,0.0004539636,0.000001256409],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001657872,"about_ca_system_score_gemma":0.0001704023,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008638726,"about_ca_topic_score_gemma":0.00003167336,"domain_scores_codex":[0.9970118,0.00008291841,0.0003086447,0.001811715,0.0001701343,0.0006148114],"domain_scores_gemma":[0.9953762,0.0002151636,0.0007984337,0.003113585,0.0002958303,0.0002008038],"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":[0.00003760473,0.00003024236,0.0002115388,0.00003466902,0.00009721989,0.0000518093,0.000473843,0.9752072,0.00000102684,0.02200054,0.00008177815,0.001772517],"study_design_scores_gemma":[0.0006078413,0.0001197526,0.000118839,0.0001449531,0.0001490792,0.00002523192,0.00007298929,0.9947014,0.000005284784,0.003436616,0.000009113763,0.0006089124],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01650597,0.00003512931,0.9808604,0.00009569692,0.0007527653,0.0009936951,0.00004700202,0.000639743,0.00006964974],"genre_scores_gemma":[0.8133399,0.00001137197,0.1858486,0.00007663408,0.0001164078,0.000004698823,0.0004725247,0.00004784537,0.00008203855],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7968339,"threshold_uncertainty_score":0.9996977,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1449076541886231,"score_gpt":0.237001323538144,"score_spread":0.09209366934952087,"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."}}