{"id":"W2948930516","doi":"10.48550/arxiv.1906.03532","title":"Reducing the variance in online optimization by transporting past gradients","year":2019,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal; University of British Columbia","funders":"","keywords":"Variance (accounting); Computer science; Econometrics; Environmental science; Statistics; Mathematics; Economics","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":[0.0003448767,0.0002893699,0.000292434,0.0002748013,0.0001218758,0.0001048531,0.001965859,0.000224289,0.00001139878],"category_scores_gemma":[0.00004918678,0.0002921745,0.0001106047,0.001011674,0.00007477464,0.0003949442,0.000609611,0.0005857287,0.000006101738],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002546002,"about_ca_system_score_gemma":0.0001079894,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001825151,"about_ca_topic_score_gemma":0.00001677949,"domain_scores_codex":[0.9979802,0.0001291989,0.0003634895,0.001047658,0.0001259644,0.000353496],"domain_scores_gemma":[0.9980863,0.0001105088,0.0004582701,0.001137455,0.0001319015,0.00007559456],"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":[0.000007395329,0.00009214297,0.002536881,0.00002670643,0.00001659351,0.0000192328,0.0003759866,0.9808432,0.000009394683,0.01572491,0.0001570248,0.0001905205],"study_design_scores_gemma":[0.0003124407,0.00002760192,0.0008758875,0.0002044573,0.00002195255,0.000002217652,0.00005057395,0.9946383,0.00003233801,0.003499553,0.00003389615,0.0003007336],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03373877,0.00004643877,0.9641868,0.0002466794,0.0005245624,0.0005882169,0.00002744012,0.0002840073,0.0003570179],"genre_scores_gemma":[0.9590862,0.0001238521,0.04011365,0.0001105687,0.00003720917,0.000002804196,0.00009987829,0.00002347377,0.0004023536],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9253474,"threshold_uncertainty_score":0.999953,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04322426133291139,"score_gpt":0.190104723853462,"score_spread":0.1468804625205506,"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."}}