{"id":"W2895642264","doi":"10.1007/978-3-030-01424-7_39","title":"Width of Minima Reached by Stochastic Gradient Descent is Influenced by Learning Rate to Batch Size Ratio","year":2018,"lang":"en","type":"book-chapter","venue":"Lecture notes in computer science","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":false,"ca_institutions":"Canadian Institute for Advanced Research; Université de Montréal","funders":"Institut de Valorisation des Données; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs; European Commission","keywords":"Maxima and minima; Stochastic gradient descent; Generalization; Computer science; Range (aeronautics); Convergence (economics); Gradient descent; Artificial neural network; Rate of convergence; Generalization error; Set (abstract data type); Artificial intelligence; Key (lock); Algorithm; Applied mathematics; Mathematics; Mathematical analysis; Materials science","routes":{"ca_aff":true,"ca_fund":true,"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":[0.001164942,0.0007426619,0.000806799,0.0007322329,0.0002771357,0.0003955522,0.003500386,0.000343514,0.00007489019],"category_scores_gemma":[0.0007695914,0.0007270421,0.0001425838,0.001080965,0.0009006379,0.0004905773,0.001351814,0.000688307,0.00003869861],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004539996,"about_ca_system_score_gemma":0.000419982,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003910029,"about_ca_topic_score_gemma":0.000008392635,"domain_scores_codex":[0.9947426,0.00008936595,0.001006098,0.002021449,0.001286443,0.0008540762],"domain_scores_gemma":[0.9957578,0.000971024,0.0007165701,0.001445611,0.0007395667,0.0003694698],"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":[0.0002155259,0.0006291287,0.000266061,0.0003535072,0.0002414588,0.00005355411,0.04015973,0.3896526,0.07318499,0.01759769,0.01475913,0.4628866],"study_design_scores_gemma":[0.0007465543,0.001993445,0.00009418443,0.001134546,0.00003945094,0.0000356413,0.000001494002,0.8859174,0.05815298,0.04910931,0.001000589,0.001774381],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.0008125589,0.0001812829,0.9959135,0.0008395359,0.0007617046,0.0009890228,0.00002021612,0.0002874828,0.0001946664],"genre_scores_gemma":[0.2568853,0.00002176897,0.7391149,0.002833438,0.0001566171,0.00006102508,0.00001744741,0.00008393528,0.0008255194],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.4962648,"threshold_uncertainty_score":0.999518,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01114941303387985,"score_gpt":0.2411353598657182,"score_spread":0.2299859468318383,"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."}}