{"id":"W2920229473","doi":"10.48550/arxiv.1903.01026","title":"Learning Modular Safe Policies in the Bandit Setting with Application to Adaptive Clinical Trials","year":2019,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Advanced Bandit Algorithms Research","field":"Decision Sciences","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Regret; Robustness (evolution); Outlier; Computer science; Modular design; Mathematical optimization; Measure (data warehouse); Estimator; Flexibility (engineering); Suite; Artificial intelligence; Machine learning; Mathematics; Data mining; Statistics","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":[0.02413842,0.001947131,0.003223144,0.001441316,0.0009103952,0.002954087,0.002062799,0.003373702,0.003276464],"category_scores_gemma":[0.07702498,0.0009075531,0.001123943,0.001659859,0.003266629,0.002992832,0.002659695,0.00498336,0.0007325808],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002685665,"about_ca_system_score_gemma":0.002684548,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002881998,"about_ca_topic_score_gemma":0.001837653,"domain_scores_codex":[0.9889756,0.008506846,0.0003522852,0.000858586,0.0008979516,0.0004086843],"domain_scores_gemma":[0.9326628,0.05835686,0.003776626,0.002416041,0.001718808,0.00106896],"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":[0.0004206043,0.0001223279,0.001357883,0.0001272217,0.0001015339,0.0000962842,0.0001151195,0.8962257,0.0002547143,0.0663416,0.001338086,0.03349892],"study_design_scores_gemma":[0.00008586403,0.00006496377,0.0001600956,0.00003318331,0.0000168126,0.00002020503,0.00001219826,0.9293809,0.0002074214,0.06945714,0.0005499162,0.00001125211],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02605385,0.001797672,0.966226,0.001994547,0.00007863305,0.0001921995,0.0001324798,0.0005034952,0.003021122],"genre_scores_gemma":[0.7207685,0.001777992,0.2711079,0.001002871,0.0002685014,0.0007788921,0.0002815919,0.0002250858,0.003788767],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.02413842,"threshold_uncertainty_score":0.1276577,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4112833942530587,"score_gpt":0.4118405645397245,"score_spread":0.0005571702866657979,"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."}}