{"id":"W6894416521","doi":"10.5555/3635637.3662926","title":"MaDi:Learning to Mask Distractions for Generalization in Visual Deep Reinforcement Learning","year":2024,"lang":"en","type":"article","venue":"TU/e Research Portal","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Generalization; Reinforcement learning; Focus (optics); Artificial neural network; Contrast (vision); Control (management); Deep 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":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.002269394,0.0001822137,0.0001794934,0.0008653722,0.0005383244,0.001119465,0.0005635716,0.0001008323,0.0001841097],"category_scores_gemma":[0.0008130202,0.0001895754,0.00009299628,0.001554663,0.0000641526,0.0007279777,0.0003851005,0.0009666692,0.0003001526],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002518249,"about_ca_system_score_gemma":0.0003128476,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001105209,"about_ca_topic_score_gemma":0.00004412564,"domain_scores_codex":[0.9967015,0.0002188026,0.0004894247,0.0006034991,0.001072029,0.0009147025],"domain_scores_gemma":[0.998758,0.0004252006,0.00006008966,0.000301042,0.000213678,0.0002420192],"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":[0.00001228665,0.00002011659,0.0009168321,0.0000785305,0.00002223536,0.00007537675,0.001657833,0.9635543,0.0009908989,0.02479473,0.0008873136,0.006989558],"study_design_scores_gemma":[0.0001986146,0.0004989714,0.0005668554,0.0001029055,0.000004158645,0.00001463062,0.0004372779,0.9536602,0.0003717415,0.0001462176,0.04379206,0.0002063963],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01128857,0.00008662503,0.9811563,0.0007544584,0.0003713761,0.00075234,4.551039e-7,0.0003156811,0.005274216],"genre_scores_gemma":[0.9791394,0.00004650291,0.006913303,0.0000407251,0.0002250509,0.000233149,0.00008985044,0.00003972379,0.01327232],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.974243,"threshold_uncertainty_score":0.9999174,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05700158475965024,"score_gpt":0.4050078222895636,"score_spread":0.3480062375299134,"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."}}