{"id":"W3166400281","doi":"10.48550/arxiv.2106.03696","title":"Dynamics of Stochastic Momentum Methods on Large-scale, Quadratic Models","year":2021,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Hessian matrix; Mathematics; Applied mathematics; Momentum (technical analysis); Hyperparameter; Quadratic equation; Stochastic optimization; Mathematical optimization; Algorithm","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0004959509,0.0003971233,0.0005964852,0.0005569864,0.0001057504,0.0000999361,0.001909623,0.0003314774,0.00002505153],"category_scores_gemma":[0.00008946428,0.000478412,0.0002828275,0.0009849333,0.0001077443,0.000361996,0.002146028,0.0005560146,0.000004848351],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004139653,"about_ca_system_score_gemma":0.0002584475,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003937543,"about_ca_topic_score_gemma":0.00002465053,"domain_scores_codex":[0.9974555,0.0003257438,0.0003919341,0.00123585,0.0001930674,0.0003978771],"domain_scores_gemma":[0.9968123,0.0002952613,0.0004883029,0.001851453,0.0003831585,0.0001695305],"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":[0.000007283796,0.0001401364,0.000007755261,0.00004972373,0.0000472897,0.00001980773,0.0003132538,0.593075,0.00000627518,0.4061633,0.00001544855,0.0001546533],"study_design_scores_gemma":[0.0002378626,0.00008345903,0.0000129389,0.0001861515,0.00006329104,0.000002337615,0.0001814763,0.8047246,0.0001526862,0.1940354,7.085472e-7,0.0003190287],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.005720102,0.00003676145,0.9915036,0.0000944772,0.0005629704,0.0004743664,0.0000339645,0.0003628321,0.0012109],"genre_scores_gemma":[0.7410585,0.00002681454,0.2583985,0.00004994079,0.00001386303,0.000003968069,0.00004632954,0.00002564792,0.0003763774],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7353384,"threshold_uncertainty_score":0.9997668,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0666013736102855,"score_gpt":0.2342714186592893,"score_spread":0.1676700450490038,"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."}}