{"id":"W2923274354","doi":"10.1016/j.neunet.2019.03.013","title":"Equivalence between dropout and data augmentation: A mathematical check","year":2019,"lang":"en","type":"article","venue":"Neural Networks","topic":"Fault Detection and Control Systems","field":"Engineering","cited_by":33,"is_retracted":false,"has_abstract":false,"ca_institutions":"Polytechnique Montréal","funders":"Southwest Petroleum University; National Natural Science Foundation of China","keywords":"Counterexample; Equivalence (formal languages); Activation function; Artificial neural network; MNIST database; Equivalence relation; Computer science; Dropout (neural networks); Representation (politics); Generalization; Nonlinear system; Quotient space (topology); Mathematics; Artificial intelligence; Machine learning; Discrete mathematics; Pure mathematics; Quotient","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":[0.00009878261,0.00009409856,0.00014885,0.00001791826,0.00002642567,0.00005398634,0.000163038,0.00006194649,0.0001187009],"category_scores_gemma":[0.000005810542,0.00008495895,0.00001688515,0.0000716967,0.00001561486,0.0001598574,0.00006291532,0.0001404424,0.00009380461],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001209437,"about_ca_system_score_gemma":0.000001726469,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003225883,"about_ca_topic_score_gemma":0.000002039826,"domain_scores_codex":[0.9993822,0.00001852804,0.0001627482,0.0001678039,0.0001053215,0.0001633748],"domain_scores_gemma":[0.9995213,0.00006277203,0.00001853547,0.0003262493,0.000006799085,0.00006436227],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001272028,0.0000737888,0.1267735,0.001810559,0.0006510395,0.00005361887,0.001068433,0.3262635,0.007011475,0.002284621,0.02803519,0.505847],"study_design_scores_gemma":[0.0003023634,0.00002125742,0.001550634,0.00002484778,0.00001258479,0.00001178347,0.00003540665,0.9965174,0.00001779406,0.00004864352,0.001359279,0.00009804367],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9650251,0.0009314161,0.02662652,0.0001939486,0.001031332,0.0004722708,0.0000156019,0.0004914145,0.005212419],"genre_scores_gemma":[0.9992074,0.00001962061,0.00007495714,0.00004679237,0.0003085187,0.000006765608,0.00001766314,0.00001650618,0.0003017544],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6702538,"threshold_uncertainty_score":0.3464525,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02491702697999136,"score_gpt":0.2682835501655859,"score_spread":0.2433665231855945,"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."}}