{"id":"W3092277419","doi":"10.1145/3424636.3426907","title":"Learning to Locomote: Understanding How Environment Design Matters for Deep Reinforcement Learning","year":2020,"lang":"en","type":"preprint","venue":"","topic":"Human Motion and Animation","field":"Engineering","cited_by":41,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Reinforcement learning; Computer science; Artificial intelligence; Animation; Action (physics); Embodied cognition; Embodied agent; Control (management); Machine learning; Human–computer interaction","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":[0.0002407503,0.0003006924,0.0002645972,0.0001444068,0.0001724975,0.0002343839,0.0001567048,0.0001476499,0.00045678],"category_scores_gemma":[0.00003188286,0.0003399909,0.0001195122,0.00004689848,0.00001119256,0.00007023351,0.0001669063,0.0005648115,0.0001823404],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008035978,"about_ca_system_score_gemma":0.000008919924,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000001856233,"about_ca_topic_score_gemma":9.647181e-7,"domain_scores_codex":[0.9988145,0.00004947337,0.0002602485,0.0003440994,0.0002240366,0.0003075892],"domain_scores_gemma":[0.9995502,0.00006133283,0.00007251215,0.0001430451,0.00001062809,0.0001623327],"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":[0.000009222361,0.000002956005,0.000003193705,0.0002963168,0.00005979663,0.000001277125,0.001769152,0.98966,0.003369026,0.0005075689,0.002716077,0.001605461],"study_design_scores_gemma":[0.0002687681,0.0001326418,0.00001195955,0.0001172797,0.00003160457,6.156887e-7,0.001811369,0.9819046,0.00158659,0.0004670245,0.01325545,0.0004121384],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0001946633,0.0000168291,0.994193,0.002592951,0.0001902482,0.001068603,6.718377e-7,0.0006176961,0.001125359],"genre_scores_gemma":[0.9527132,0.00006644473,0.04433477,0.0004300087,0.0001516792,0.000195134,0.0001166136,0.0001098744,0.001882285],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9525185,"threshold_uncertainty_score":0.9999052,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06800566205986544,"score_gpt":0.2325204118775476,"score_spread":0.1645147498176821,"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."}}