{"id":"W3034736247","doi":"10.24963/ijcai.2020/374","title":"SVRG for Policy Evaluation with Fewer Gradient Evaluations","year":2020,"lang":"en","type":"preprint","venue":"","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal; McGill University","funders":"","keywords":"Reinforcement learning; Variance (accounting); Computer science; Convergence (economics); Function (biology); Artificial intelligence; Computation; Mathematical optimization; Gradient method; Bellman equation; Work (physics); Scale (ratio); Value (mathematics); Machine learning; Mathematics; 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":[0.0007831723,0.000334991,0.0003154128,0.0004450176,0.0001536049,0.0003286213,0.001212014,0.0001751388,0.00004251497],"category_scores_gemma":[0.000707517,0.0002863286,0.0001390511,0.0006393028,0.00006159938,0.0002197489,0.000771865,0.0002367612,0.00002427588],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004692997,"about_ca_system_score_gemma":0.00168266,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005683449,"about_ca_topic_score_gemma":0.00001187424,"domain_scores_codex":[0.9971249,0.0001137354,0.0004157753,0.001007083,0.001021554,0.0003169377],"domain_scores_gemma":[0.9968136,0.0001549591,0.0003238048,0.0011271,0.001408908,0.0001716507],"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":[0.00001660088,0.000100704,0.00002355252,0.00007637723,0.0001059495,6.857324e-7,0.0009912469,0.06618708,0.00003147574,0.8976206,0.008708293,0.02613746],"study_design_scores_gemma":[0.000424498,0.0002082023,0.0001600348,0.00005704042,0.00007848739,0.000003701588,0.000009475808,0.7536902,0.0004674776,0.2441598,0.0004434724,0.0002975867],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00008458542,0.0000458152,0.9731936,0.01689745,0.0004201157,0.004155879,0.00002662022,0.0009516062,0.004224343],"genre_scores_gemma":[0.1634942,0.00000865942,0.8313327,0.000987599,0.0002480184,0.003395697,0.0002107706,0.00003928316,0.0002830862],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.6875031,"threshold_uncertainty_score":0.9999589,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08629240404755965,"score_gpt":0.3720355181123799,"score_spread":0.2857431140648203,"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."}}