{"id":"W2966312692","doi":"","title":"Normal Approximation for Stochastic Gradient Descent via Non-Asymptotic Rates of Martingale CLT","year":2019,"lang":"en","type":"article","venue":"Conference on Learning Theory","topic":"Markov Chains and Monte Carlo Methods","field":"Mathematics","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Mathematics; Martingale (probability theory); Martingale difference sequence; Central limit theorem; Applied mathematics; Stochastic gradient descent; Rate of convergence; Multivariate random variable; Weak convergence; Mathematical analysis; Random variable; Statistics; Computer science","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":[0.01918491,0.001971575,0.002046724,0.00330836,0.0008111699,0.002340034,0.003037557,0.001955783,0.004485499],"category_scores_gemma":[0.08012067,0.0009018084,0.002311732,0.001442717,0.00488939,0.004796594,0.004458407,0.005467314,0.0008278421],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003232388,"about_ca_system_score_gemma":0.002412873,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003627175,"about_ca_topic_score_gemma":0.002608439,"domain_scores_codex":[0.9947425,0.002917537,0.0002316536,0.0005952699,0.001155854,0.0003570772],"domain_scores_gemma":[0.9553038,0.03328168,0.002354759,0.002515396,0.004931005,0.001613319],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001198965,0.00007124269,0.002050213,0.0002593969,0.00008876423,0.0002160148,0.0002396051,0.1538798,0.002052334,0.8218493,0.001375466,0.01779811],"study_design_scores_gemma":[0.00001304437,0.00004547257,0.0003290665,0.00004293922,0.00001550761,0.00005920382,0.00001519474,0.850823,0.0009122188,0.1469517,0.0007665451,0.00002610067],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01042259,0.0005532646,0.985599,0.0004536083,0.0000936785,0.0000576621,0.0000515263,0.0002080397,0.002560546],"genre_scores_gemma":[0.6049209,0.001970904,0.3763709,0.0008891462,0.0004638174,0.00100657,0.0004715222,0.0007497591,0.01315641],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01918491,"threshold_uncertainty_score":0.1014607,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05635044454367171,"score_gpt":0.3348542065146358,"score_spread":0.278503761970964,"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."}}