{"id":"W4386185082","doi":"10.48550/arxiv.2308.12445","title":"An Intentional Forgetting-Driven Self-Healing Method For Deep Reinforcement Learning Systems","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Forgetting; Reinforcement learning; Convergence (economics); Adaptation (eye); Computer science; Reinforcement; Artificial intelligence; Cognitive psychology; Psychology; Social psychology; Neuroscience","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":[],"consensus_categories":[],"category_scores_codex":[0.001302703,0.0007607979,0.0007496875,0.0003843755,0.0003819843,0.0005655109,0.001474224,0.0007376531,0.001966628],"category_scores_gemma":[0.002298406,0.0003478724,0.0005259843,0.0002085651,0.0006834363,0.0007030955,0.001276475,0.001265256,0.0002858982],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007183336,"about_ca_system_score_gemma":0.0009254668,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003337622,"about_ca_topic_score_gemma":0.003331404,"domain_scores_codex":[0.9995493,0.0001111975,0.00003530599,0.00009861363,0.0001288317,0.00007680784],"domain_scores_gemma":[0.9991205,0.0003473998,0.0001375025,0.0001095291,0.0002041548,0.00008097916],"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":[0.0001104243,0.00009264528,0.001314811,0.0001203362,0.00006593963,0.0001070279,0.0001453401,0.8342296,0.007400961,0.008043576,0.001563328,0.146806],"study_design_scores_gemma":[0.000008112584,0.00002797519,0.00004390702,0.000003477379,0.000004687079,0.00001003839,0.0000035539,0.9981358,0.0004823661,0.001003419,0.0002737056,0.000002932657],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02686524,0.0003484673,0.9691513,0.0002085308,0.00006827913,0.00006494149,0.00002317014,0.00125496,0.002015021],"genre_scores_gemma":[0.8853337,0.0001374435,0.1113843,0.0002388484,0.00004839278,0.0001474462,0.00005932442,0.0001139899,0.002536396],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003337622,"threshold_uncertainty_score":0.006889462,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08574056855884084,"score_gpt":0.2481800766699188,"score_spread":0.1624395081110779,"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."}}