{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00211507,0.0001882598,0.0003563934,0.000104762,0.00008880832,0.00002559866,0.0001777562,0.00009182743,0.00010866],"category_scores_gemma":[0.001289536,0.0001645153,0.0001265175,0.00008129331,0.00006528571,0.00006267161,0.00005330564,0.000240134,0.000002937158],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004373513,"about_ca_system_score_gemma":0.00005557476,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000002459884,"about_ca_topic_score_gemma":0.000001454462,"domain_scores_codex":[0.9986227,0.0002670644,0.0003618591,0.0002705721,0.0001837057,0.0002941112],"domain_scores_gemma":[0.9976437,0.00142475,0.0003507704,0.0003002552,0.0002099674,0.00007057019],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001021464,0.0003940159,0.001335946,0.001512261,0.0001136376,0.000001325341,0.005416746,0.003198275,0.03304805,0.898644,0.0001656865,0.05514863],"study_design_scores_gemma":[0.004071291,0.003243958,0.001033364,0.001833092,0.0002452432,0.00001259604,0.005334501,0.6888132,0.03683381,0.2570305,0.0004389365,0.001109458],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.422318,0.00001318737,0.5730914,0.00003130175,0.0001955787,0.0005926869,0.000002869436,0.00004116303,0.003713735],"genre_scores_gemma":[0.9816867,0.000002981856,0.01502221,0.00002737559,0.00004805639,0.00006639253,0.00001226075,0.00003456151,0.00309947],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6856149,"threshold_uncertainty_score":0.670874,"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."}}