{"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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001165856,0.0004913879,0.0005464244,0.0005726917,0.0005783328,0.0005151475,0.002605955,0.0004256863,0.000008851622],"category_scores_gemma":[0.0001869399,0.00060139,0.0003891685,0.0005952626,0.00004645041,0.0006524299,0.002058689,0.001030694,0.00007919655],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000623844,"about_ca_system_score_gemma":0.0002714974,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001635891,"about_ca_topic_score_gemma":0.000009326922,"domain_scores_codex":[0.9964191,0.0003795134,0.0005996358,0.001528994,0.0003000924,0.0007726168],"domain_scores_gemma":[0.9964753,0.0005151676,0.0008596616,0.001308054,0.0005578719,0.0002839011],"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":[0.00001655334,0.00002460488,0.0008599635,0.0003295614,0.0002301211,0.00003792614,0.0005522502,0.8797761,0.00002403588,0.1179703,0.00008219687,0.00009644854],"study_design_scores_gemma":[0.0005738736,0.000245586,0.00008797117,0.0002294294,0.0001329711,0.000004833689,0.0004711211,0.9946494,0.00002221509,0.001990343,0.0009751945,0.0006170762],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0009958767,0.00002727904,0.9938477,0.00006954416,0.001986084,0.001074928,0.000002999072,0.001548549,0.000447019],"genre_scores_gemma":[0.8461806,0.00007423031,0.1481124,0.0000602069,0.0002718535,0.00001909072,0.0002131249,0.00008159578,0.004986956],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8457353,"threshold_uncertainty_score":0.9996437,"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."}}