{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002646167,0.001082498,0.0009389324,0.0006428267,0.0006162978,0.001288076,0.001422091,0.001485751,0.001892587],"category_scores_gemma":[0.01269308,0.000631635,0.0005792348,0.0005332761,0.002250362,0.002086171,0.002049346,0.00180545,0.0003655627],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001380012,"about_ca_system_score_gemma":0.001135835,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004286844,"about_ca_topic_score_gemma":0.003548826,"domain_scores_codex":[0.9991498,0.0004036771,0.0000254569,0.000116027,0.0002236705,0.0000812764],"domain_scores_gemma":[0.9962012,0.002642379,0.0004427302,0.0002084511,0.000304562,0.0002006735],"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":[0.00005506126,0.00003245973,0.0005746814,0.00005374044,0.00003237752,0.00007258832,0.00006465258,0.8793884,0.001171456,0.1091641,0.001177764,0.008212801],"study_design_scores_gemma":[0.000003610465,0.000006204551,0.00003703824,0.000002767814,9.653396e-7,0.000003804464,0.000002021727,0.9886573,0.00007414069,0.01108591,0.0001240245,0.000002164568],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02738378,0.0002870962,0.9686612,0.0007588999,0.00004301729,0.0000342861,0.00003318835,0.0002102385,0.002588371],"genre_scores_gemma":[0.7966905,0.0006520061,0.1907795,0.0004025921,0.0001834331,0.0003296878,0.0001989232,0.0002989044,0.01046444],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004286844,"threshold_uncertainty_score":0.0139944,"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."}}