{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003336208,0.00100415,0.001330144,0.0007357082,0.0003106146,0.001240356,0.002866333,0.001349091,0.002313942],"category_scores_gemma":[0.007842985,0.0005614876,0.0007221775,0.0006742583,0.0009959224,0.002450126,0.001948342,0.002768209,0.0009701276],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001066409,"about_ca_system_score_gemma":0.001614053,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005704152,"about_ca_topic_score_gemma":0.005444581,"domain_scores_codex":[0.9985805,0.000385061,0.0001194713,0.0002969338,0.0005158507,0.0001022583],"domain_scores_gemma":[0.9972796,0.0009452719,0.0002709005,0.0007555078,0.0006099512,0.0001387341],"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":[0.0002381711,0.0002254025,0.001554183,0.00008785429,0.0001103335,0.000126453,0.00006236168,0.7809908,0.006860259,0.02251605,0.00539534,0.1818327],"study_design_scores_gemma":[0.00001029046,0.00001730956,0.00003687428,0.000003239626,0.000002873573,0.000006862888,0.000001529782,0.9963258,0.0007628744,0.002616125,0.0002135499,0.000002660976],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05774147,0.0006448878,0.9354084,0.0005853649,0.0002008909,0.00007546158,0.0001111538,0.003317852,0.001914501],"genre_scores_gemma":[0.6374775,0.0003267146,0.355829,0.0007030408,0.0001209831,0.0001583299,0.0004376985,0.0004175337,0.004529302],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005704152,"threshold_uncertainty_score":0.01764375,"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."}}