{"id":"W4226334043","doi":"10.1016/j.ipm.2022.102933","title":"ARL: An adaptive reinforcement learning framework for complex question answering over knowledge base","year":2022,"lang":"en","type":"article","venue":"Information Processing & Management","topic":"Topic Modeling","field":"Computer Science","cited_by":55,"is_retracted":false,"has_abstract":false,"ca_institutions":"York University","funders":"","keywords":"Interpretability; Computer science; Reinforcement learning; Artificial intelligence; Knowledge base; Benchmark (surveying); Metric (unit); Machine learning; Relation (database); Question answering; Data mining","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":[],"consensus_categories":[],"category_scores_codex":[0.0009190507,0.0001547554,0.0001241668,0.0002800814,0.000907207,0.000458562,0.0005756888,0.00003121045,0.00004186417],"category_scores_gemma":[0.00002899566,0.0001761389,0.00004474498,0.0003731724,0.00001440005,0.003387862,0.0005592997,0.000221471,0.00001313455],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003348338,"about_ca_system_score_gemma":0.00005622267,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001246448,"about_ca_topic_score_gemma":0.000001259972,"domain_scores_codex":[0.9984772,0.00006116046,0.0004601935,0.0002503321,0.0004287003,0.0003224025],"domain_scores_gemma":[0.9991332,0.00003259264,0.0003095098,0.0003248264,0.0001256853,0.00007418491],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002005152,0.00002522591,0.00001527644,0.0001455656,0.00001075212,6.871949e-7,0.004806161,0.38656,0.000004502102,0.1889741,0.0001026657,0.419335],"study_design_scores_gemma":[0.0003818372,0.0001461113,0.0002689206,0.00007028747,0.00001043754,0.000002115188,0.00148889,0.955884,0.00002057547,0.007674823,0.03384358,0.0002084728],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.001432929,0.00002929035,0.9924501,0.0001947117,0.0002852821,0.000568944,0.000001275126,0.0003831841,0.004654308],"genre_scores_gemma":[0.7114611,0.000002642577,0.2873932,0.0005438424,0.00004228914,0.0003169022,0.0000471997,0.00000838374,0.0001844889],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7100282,"threshold_uncertainty_score":0.7182735,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03806807387468865,"score_gpt":0.3012217434974285,"score_spread":0.2631536696227399,"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."}}