{"id":"W4308071056","doi":"10.48550/arxiv.2110.01954","title":"Continuous-Time Fitted Value Iteration for Robust Policies","year":2021,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"European Commission; Canadian Institute for Advanced Research; Nvidia","keywords":"Bellman equation; Discretization; Reinforcement learning; Leverage (statistics); Mathematical optimization; Optimal control; Computer science; Dynamic programming; Robustness (evolution); Mathematics; Artificial intelligence","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0003540723,0.0003823985,0.0004906261,0.0002989771,0.0003187656,0.0005273836,0.001754508,0.0003833679,0.00004692271],"category_scores_gemma":[0.0002741355,0.000478078,0.0003229407,0.0005453185,0.0000895433,0.0007633932,0.002259996,0.00063015,0.0000416419],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002567324,"about_ca_system_score_gemma":0.0002875267,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002297376,"about_ca_topic_score_gemma":0.00001880895,"domain_scores_codex":[0.9977126,0.0002191463,0.0002677855,0.001233374,0.000123897,0.0004432689],"domain_scores_gemma":[0.9976354,0.0002985336,0.0003830715,0.001176891,0.0003965377,0.0001095128],"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":[0.00001835105,0.00004395955,0.0005420362,0.00007084325,0.00007876382,0.00007740913,0.0003851227,0.9143702,0.00008860903,0.08359119,0.0003705439,0.0003629311],"study_design_scores_gemma":[0.0005687817,0.00003932865,0.0004708531,0.0001256255,0.00007983938,0.000005491534,0.00008904532,0.9899741,0.0001310735,0.007423574,0.0005919459,0.0005003865],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06473346,0.00004017433,0.9318313,0.0004099957,0.0008732715,0.0004616614,0.00001328892,0.0004126908,0.001224156],"genre_scores_gemma":[0.9384173,0.00002418382,0.05687527,0.0003351336,0.0003525376,0.000003940595,0.0001470634,0.00004195661,0.003802615],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.874956,"threshold_uncertainty_score":0.9997671,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05323358205397503,"score_gpt":0.2011440867758975,"score_spread":0.1479105047219225,"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."}}