{"id":"W7043660303","doi":"","title":"Stochastic variance reduced gradient for policy evaluation with fewer gradient evaluations","year":2020,"lang":"en","type":"dissertation","venue":"eScholarship@McGill (McGill)","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Reinforcement learning; Convergence (economics); Gradient method; Computation; Variance (accounting); Function (biology); Gradient descent; Bellman equation","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","sts"],"consensus_categories":[],"category_scores_codex":[0.001876562,0.00120346,0.001020971,0.001274533,0.001646609,0.0003822192,0.002207231,0.0006312145,0.00005534541],"category_scores_gemma":[0.004154064,0.001231593,0.0004333307,0.002739179,0.0001071139,0.00137927,0.0002322667,0.001028648,0.0001092873],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002575713,"about_ca_system_score_gemma":0.0009608583,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009349276,"about_ca_topic_score_gemma":0.0001509029,"domain_scores_codex":[0.9917793,0.0005331743,0.001385675,0.002557154,0.002590999,0.001153686],"domain_scores_gemma":[0.9923981,0.0005328021,0.001429497,0.001816984,0.003148396,0.000674193],"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":[0.0002487904,0.0003301601,6.752333e-7,0.0002113084,0.0003183471,0.000008374286,0.0001248528,0.01078141,0.003685088,0.8895034,0.00004464806,0.09474296],"study_design_scores_gemma":[0.00699126,0.003542481,0.0005966498,0.002040139,0.002059336,0.0001392949,0.0002719876,0.347596,0.03629793,0.5878988,0.00749859,0.005067578],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1321829,0.001495157,0.6885185,0.004273161,0.0184924,0.06342984,0.005599986,0.01137378,0.07463428],"genre_scores_gemma":[0.86518,0.00003566711,0.1225044,0.0005844402,0.0001974916,0.006641769,0.002787556,0.0003480001,0.001720758],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.732997,"threshold_uncertainty_score":0.9996531,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03332380938638713,"score_gpt":0.3014796231985526,"score_spread":0.2681558138121655,"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."}}