{"id":"W3012024403","doi":"10.1162/neco_a_01276","title":"The Stochastic Delta Rule: Faster and More Accurate Deep Learning Through Adaptive Weight Noise","year":2020,"lang":"en","type":"article","venue":"Neural Computation","topic":"Neural Networks and Applications","field":"Computer Science","cited_by":11,"is_retracted":false,"has_abstract":true,"ca_institutions":"Baycrest Hospital","funders":"","keywords":"Overfitting; Dropout (neural networks); Computer science; Artificial neural network; Bayes' theorem; Algorithm; Benchmark (surveying); Random variable; Standard deviation; Noise (video); Feature (linguistics); Artificial intelligence; Pattern recognition (psychology); Mathematics; Machine learning; Bayesian probability; Statistics; Image (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":[],"consensus_categories":[],"category_scores_codex":[0.00004814899,0.0001295012,0.0001013377,0.00001227321,0.0005067213,0.0002798383,0.0003020905,0.00003004933,0.000001300602],"category_scores_gemma":[0.00002155231,0.00009314566,0.00003271429,0.0003049728,0.00006155154,0.000476211,0.0002087449,0.0002333032,0.00001766788],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001090963,"about_ca_system_score_gemma":0.00001015827,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000007994821,"about_ca_topic_score_gemma":0.000003195604,"domain_scores_codex":[0.9990466,0.0000763607,0.0001765093,0.0003303698,0.0001723726,0.0001978071],"domain_scores_gemma":[0.9993417,0.000272038,0.0001122094,0.0001168703,0.00007495667,0.00008217955],"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":[0.00003442437,0.00001550924,0.00009531125,0.000008464556,0.00001802195,0.000008305752,0.003555453,0.6728553,0.0004331376,0.01299472,0.0006196026,0.3093617],"study_design_scores_gemma":[0.0001812434,0.00008462601,0.002123989,0.000005311678,0.000006972699,0.00001000945,0.00009322399,0.9944768,0.00003480063,0.002348911,0.0005237481,0.0001103473],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0712861,0.000222825,0.9168825,0.01104981,0.0001077696,0.0002200774,0.0000012597,0.0001410899,0.00008859039],"genre_scores_gemma":[0.9931526,0.00001759405,0.005167576,0.001463709,0.0001395346,0.00002142514,0.000007121529,0.000009750832,0.00002070061],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9218665,"threshold_uncertainty_score":0.3897342,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02878424413423607,"score_gpt":0.2667397132608819,"score_spread":0.2379554691266458,"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."}}