{"id":"W2786615446","doi":"10.48550/arxiv.1803.02021","title":"Understanding Short-Horizon Bias in Stochastic Meta-Optimization","year":2018,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":45,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Horizon; Hyperparameter; Computer science; Benchmark (surveying); Time horizon; Artificial neural network; Optimization problem; Meta learning (computer science); Mathematical optimization; Machine learning; Algorithm; Mathematics; Economics","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":[0.0002356507,0.0003251634,0.0003976727,0.0004072789,0.0001678438,0.00009971832,0.00143729,0.0002286542,0.0000339571],"category_scores_gemma":[0.00003930835,0.0003680921,0.0001883677,0.001314035,0.0001271036,0.0004466977,0.0014362,0.0004897943,0.00004151247],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006539043,"about_ca_system_score_gemma":0.0001034171,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002062122,"about_ca_topic_score_gemma":0.00007156659,"domain_scores_codex":[0.9979099,0.000137742,0.0002581472,0.001212561,0.0001116162,0.0003700193],"domain_scores_gemma":[0.9981478,0.000217599,0.0002086874,0.001195866,0.0001001957,0.0001298058],"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":[0.000007301172,0.00003157696,0.00007152478,0.000007881702,0.0000814192,0.0000273134,0.00003663596,0.9013765,0.000003571837,0.0981904,0.0001122334,0.0000536753],"study_design_scores_gemma":[0.0001327748,0.00003902005,0.00002529302,0.00003073723,0.0001685979,0.000002195949,0.00002106984,0.9149222,0.00001744537,0.0842874,0.00003160244,0.00032163],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.004643966,0.00004345205,0.9932688,0.00011441,0.0003732849,0.0004648725,0.000008966994,0.0002823998,0.0007998301],"genre_scores_gemma":[0.9831505,0.000081879,0.01641902,0.00004383596,0.0000765723,0.00000584771,0.00002146668,0.00002348634,0.0001773362],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9785066,"threshold_uncertainty_score":0.9998771,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4050751399677809,"score_gpt":0.2476757384022235,"score_spread":0.1573994015655573,"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."}}