{"id":"W2948937890","doi":"10.48550/arxiv.1903.03176","title":"MinAtar: An Atari-Inspired Testbed for Thorough and Reproducible Reinforcement Learning Experiments","year":2019,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":42,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Reinforcement learning; Testbed; Computer science; Representation (politics); Breakout; Artificial intelligence; Set (abstract data type); Human–computer interaction; Toolbox; Machine learning","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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.002048306,0.001469242,0.0008491393,0.0004980557,0.0003447204,0.0007651753,0.003603885,0.001229447,0.01058409],"category_scores_gemma":[0.005764442,0.0004383112,0.0006705936,0.0004087429,0.0007107354,0.001121054,0.001068439,0.002165598,0.002095865],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000662479,"about_ca_system_score_gemma":0.0007820145,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002469925,"about_ca_topic_score_gemma":0.003200828,"domain_scores_codex":[0.9990733,0.0003849098,0.00005764649,0.0001871755,0.0001838832,0.0001130585],"domain_scores_gemma":[0.9973534,0.001484876,0.0001600944,0.0005070961,0.0002481545,0.0002463259],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":"reproducibility","domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.003038386,0.003482374,0.006125166,0.002090057,0.0005546996,0.0008286452,0.0004482786,0.7718283,0.03448028,0.04351679,0.05173916,0.08186785],"study_design_scores_gemma":[0.0005495095,0.0008443117,0.001547942,0.00007708123,0.00004766905,0.0001162399,0.00008275628,0.944185,0.01613409,0.01689886,0.01946512,0.00005139224],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"software","genre_scores_codex":[0.3636325,0.001461418,0.5438235,0.001267521,0.0009524352,0.001500497,0.01276325,0.0417809,0.03281798],"genre_scores_gemma":[0.6322874,0.0003919254,0.3489139,0.0004482165,0.00004601274,0.001714252,0.007849587,0.002494227,0.005854496],"genre_candidate":"software","genre_consensus":null,"teacher_disagreement_score":0.9979517,"threshold_uncertainty_score":0.0354073,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1109329489629851,"score_gpt":0.2332999492194836,"score_spread":0.1223670002564985,"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."}}