{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002253953,0.0007471623,0.001036478,0.0007088141,0.0004154245,0.001223576,0.00340411,0.001409805,0.00848664],"category_scores_gemma":[0.005932056,0.0006166711,0.0008977554,0.0005369014,0.0007072294,0.002051875,0.002092668,0.002504505,0.002231376],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009485146,"about_ca_system_score_gemma":0.001272143,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008216849,"about_ca_topic_score_gemma":0.01165016,"domain_scores_codex":[0.9991291,0.0003436772,0.00004634295,0.0001915385,0.0002057984,0.00008352812],"domain_scores_gemma":[0.9982848,0.001060062,0.00008760417,0.000201194,0.0002476737,0.0001187617],"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":[0.0005158728,0.00052603,0.001158614,0.0003918624,0.0001725506,0.0002007033,0.0003509349,0.4500276,0.007402636,0.03224195,0.0176263,0.489385],"study_design_scores_gemma":[0.00002699713,0.00002375566,0.00004922986,0.000007105842,0.00001010351,0.000009950953,0.000008611914,0.9889051,0.0006485434,0.008557246,0.001746408,0.000006968981],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003173174,0.0001410713,0.9893959,0.0001531161,0.0000388738,0.00008921047,0.0001823578,0.00599973,0.0008266856],"genre_scores_gemma":[0.2021127,0.0002323265,0.7903991,0.0003370952,0.00009015576,0.0004521044,0.0008121919,0.0006992914,0.004864949],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.00848664,"threshold_uncertainty_score":0.02839059,"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."}}