{"id":"W4315881264","doi":"10.48550/arxiv.2301.04268","title":"Adversarial Online Multi-Task Reinforcement Learning","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Machine Learning and Algorithms","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Omega; Combinatorics; Sample complexity; Regret; Reinforcement learning; Upper and lower bounds; Lambda; Mathematics; Tilde; Discrete mathematics; Finite set; Artificial intelligence; Computer science; Physics; Statistics","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0003799984,0.0003874713,0.0003779708,0.0003602466,0.0003008065,0.0001604594,0.002013253,0.0002971616,0.00003662222],"category_scores_gemma":[0.0001375253,0.0004460892,0.0002840417,0.0006193346,0.00006971804,0.0002451281,0.003788554,0.001652403,0.0004495222],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000194071,"about_ca_system_score_gemma":0.0002034285,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008645176,"about_ca_topic_score_gemma":0.00005894433,"domain_scores_codex":[0.9975396,0.0002239004,0.0002702271,0.001293038,0.0001689895,0.000504183],"domain_scores_gemma":[0.9981406,0.000127566,0.0003341319,0.001059515,0.0001268586,0.000211281],"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":[0.00001294326,0.00006406334,0.001309452,0.00004465666,0.00008818134,0.0003626688,0.0003125196,0.9869823,0.00001090564,0.008966384,0.0002455417,0.001600375],"study_design_scores_gemma":[0.0008461553,0.00008158368,0.001194435,0.00009653575,0.00004274883,0.000002648145,0.00008042388,0.9919767,0.000010005,0.001541285,0.003643359,0.0004841278],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02431276,0.00002338904,0.9712306,0.0002408407,0.002031292,0.000240199,0.000009378792,0.001259542,0.0006519561],"genre_scores_gemma":[0.9588168,0.0001477947,0.007992747,0.00008484107,0.0003444357,0.000001020584,0.0001407667,0.00003772706,0.03243386],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9632379,"threshold_uncertainty_score":0.9997991,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09278095900279544,"score_gpt":0.2197380260531538,"score_spread":0.1269570670503584,"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."}}