{"id":"W7015986514","doi":"","title":"Towards better generalization capabilities of reinforcement learning agents via self-supervision","year":2023,"lang":"en","type":"dissertation","venue":"eScholarship@McGill (McGill)","topic":"Reinforcement Learning in Robotics","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Generalization; Reinforcement learning; Control (management); Feature (linguistics); Stability (learning theory)","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":[0.00138006,0.0009106017,0.0009255773,0.000882609,0.001015583,0.0002377639,0.002058853,0.0008119142,0.0002188288],"category_scores_gemma":[0.0006752349,0.0009868466,0.0004736052,0.001283489,0.00005245108,0.001494034,0.0006939099,0.001439742,0.0004216849],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007057918,"about_ca_system_score_gemma":0.0001186426,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005429303,"about_ca_topic_score_gemma":0.0000668116,"domain_scores_codex":[0.9931929,0.000549521,0.001775768,0.001329631,0.002208137,0.0009440547],"domain_scores_gemma":[0.9959068,0.0002040208,0.001285758,0.001446703,0.0008794699,0.0002772462],"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":[0.00006351413,0.0001275402,0.0001791493,0.002199822,0.00070799,0.00004117021,0.0005390967,0.8325425,0.01137146,0.05590975,0.00004034063,0.09627764],"study_design_scores_gemma":[0.002851941,0.001727873,0.004568593,0.002266666,0.0007935581,0.00002928037,0.0009427285,0.7694554,0.1263541,0.01259381,0.07385091,0.00456513],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9122621,0.000246144,0.008958535,0.00008071905,0.008928328,0.002891677,0.0000842281,0.003610941,0.0629373],"genre_scores_gemma":[0.9571602,0.0004792578,0.01247734,0.0002739212,0.00009525456,0.0001605701,0.001869668,0.0002980053,0.02718581],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1149826,"threshold_uncertainty_score":0.9992582,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01924933637974323,"score_gpt":0.2519997002253784,"score_spread":0.2327503638456352,"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."}}