{"id":"W4376134510","doi":"10.1137/22m1530070","title":"Beating a Benchmark: Dynamic Programming May Not Be the Right Numerical Approach","year":2023,"lang":"en","type":"article","venue":"SIAM Journal on Financial Mathematics","topic":"Financial Markets and Investment Strategies","field":"Economics, Econometrics and Finance","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"","keywords":"Benchmark (surveying); Dynamic programming; Mathematical optimization; Reinforcement learning; Bellman equation; Optimal control; Computer science; Stochastic control; Optimization problem; Mathematics; Artificial intelligence","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":[0.001885292,0.0003327024,0.0006279856,0.000302124,0.0008628807,0.0004056912,0.000573682,0.0001715697,0.0001608303],"category_scores_gemma":[0.0008774914,0.0002512122,0.0003258699,0.0008620217,0.0001444476,0.0002767456,0.0001096387,0.0007300891,0.0003749622],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001490917,"about_ca_system_score_gemma":0.00008972552,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000008305924,"about_ca_topic_score_gemma":0.000002292441,"domain_scores_codex":[0.9974959,0.00004146474,0.00115359,0.0003738139,0.0002153852,0.0007198437],"domain_scores_gemma":[0.9983382,0.0002482919,0.0008309339,0.0003956873,0.00005474846,0.0001321888],"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":[0.00002389583,0.0002713321,0.0002674928,0.0001130218,0.00003942662,0.00004177889,0.001453759,0.000223384,0.000008483574,0.9884643,0.005464214,0.00362892],"study_design_scores_gemma":[0.001551215,0.0009926406,0.0280255,0.0004062541,0.00006064053,0.0002987612,0.001416374,0.05788938,0.00006014858,0.4796915,0.4281068,0.001500811],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.604524,0.00185862,0.03618874,0.02803742,0.006001168,0.00314878,0.0003803773,0.000772248,0.3190886],"genre_scores_gemma":[0.9772409,0.0003725387,0.01761105,0.001262427,0.000645,0.0001127372,0.00002872644,0.00009002389,0.002636667],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5087728,"threshold_uncertainty_score":0.999994,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03803374853306173,"score_gpt":0.2483354425064158,"score_spread":0.210301693973354,"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."}}