{"id":"W7124241732","doi":"10.65109/xgyf2237","title":"Interpretable Preference-based Reinforcement Learning with Tree-Structured Reward Functions","year":2022,"lang":"","type":"article","venue":"","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Thales (Canada)","funders":"","keywords":"Interpretability; Reinforcement learning; Robustness (evolution); Function (biology); Heuristic; Preference","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","sts","research_integrity","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.000968532,0.0008055867,0.0006769804,0.0006345236,0.002982613,0.0009948167,0.002435041,0.0001413214,0.01569544],"category_scores_gemma":[0.0001376419,0.0007470221,0.000276205,0.002222819,0.0002536535,0.0009144087,0.002136319,0.002443864,0.0001823416],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001009454,"about_ca_system_score_gemma":0.001425323,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000212655,"about_ca_topic_score_gemma":0.00003998445,"domain_scores_codex":[0.9928274,0.0006613047,0.001173961,0.001488451,0.002399022,0.001449857],"domain_scores_gemma":[0.9962199,0.0002638127,0.0008869416,0.001861338,0.0003504819,0.0004175655],"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":[0.0003556256,0.00007184241,0.002764145,0.00005984546,0.0002165331,0.00002875802,0.001816236,0.9808074,0.0001263956,0.002213994,0.002541288,0.008997927],"study_design_scores_gemma":[0.001984044,0.005468548,0.0003351801,0.00009918812,0.0001401205,0.00004298189,0.001066203,0.9193419,0.0003357382,0.00002842985,0.07024059,0.0009170944],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.001347855,0.00008084009,0.9207791,0.0008600441,0.001780733,0.001038807,0.000003701381,0.0005877181,0.07352117],"genre_scores_gemma":[0.8423514,0.00001130694,0.01461309,0.0008313879,0.0000876833,0.0002453893,0.00009346323,0.00006619223,0.1417],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.906166,"threshold_uncertainty_score":0.9998575,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02839159268205176,"score_gpt":0.2199169094736677,"score_spread":0.1915253167916159,"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."}}