{"id":"W4386566626","doi":"10.18653/v1/2023.eacl-main.88","title":"Policy-based Reinforcement Learning for Generalisation in Interactive Text-based Environments","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"Institut National de la Recherche Scientifique","funders":"University of Cape Town; National Research Foundation","keywords":"Converse; Reinforcement learning; Computer science; Benchmark (surveying); Natural language; Artificial intelligence; Variety (cybernetics); Simple (philosophy); Question answering; Value (mathematics); Natural language understanding; Baseline (sea); Machine learning; Mathematics","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":[0.002773597,0.00107587,0.00123154,0.0005540589,0.00049539,0.00110012,0.002016339,0.001326005,0.002433119],"category_scores_gemma":[0.01338793,0.0004776539,0.000546902,0.0003981184,0.001458513,0.002390011,0.00164756,0.002136103,0.0006271133],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00110748,"about_ca_system_score_gemma":0.0009583036,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005149796,"about_ca_topic_score_gemma":0.004898888,"domain_scores_codex":[0.9987463,0.0006235559,0.00005725381,0.0003065121,0.000153021,0.0001134272],"domain_scores_gemma":[0.9942656,0.004519222,0.00024425,0.0004336592,0.0003436525,0.0001935847],"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":[0.0002252288,0.0002140996,0.001415498,0.0001544673,0.00005992802,0.0001504246,0.0004054318,0.8791914,0.005983117,0.00897632,0.001517317,0.1017068],"study_design_scores_gemma":[0.00001814268,0.00003621547,0.00009069669,0.000005824892,0.000005058023,0.00001127516,0.00001997646,0.9937041,0.0008702619,0.004892911,0.0003386883,0.000006775861],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.06008566,0.0002733141,0.9338612,0.0003178331,0.00005500729,0.0001311067,0.00007830976,0.00284162,0.002356062],"genre_scores_gemma":[0.8563341,0.0001364623,0.1402325,0.0002290666,0.00004424822,0.0002497914,0.0002122564,0.0002921486,0.002269414],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005149796,"threshold_uncertainty_score":0.01466835,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.035474172160724,"score_gpt":0.2884375952942587,"score_spread":0.2529634231335347,"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."}}