{"id":"W2591984255","doi":"10.17863/cam.21343","title":"Sequence Tutor: Conservative Fine-Tuning of Sequence Generation Models with KL-control","year":2017,"lang":"en","type":"article","venue":"Apollo (University of Cambridge)","topic":"Music and Audio Processing","field":"Computer Science","cited_by":76,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal; Google (Canada)","funders":"Engineering and Physical Sciences Research Council; Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Recurrent neural network; Sequence (biology); Reinforcement learning; Security token; Artificial intelligence; Machine learning; Artificial neural network","routes":{"ca_aff":true,"ca_fund":true,"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":[0.0002102837,0.0001220316,0.0002548066,0.00007368223,0.0005437833,0.0000741439,0.001027415,0.00005347835,0.000004792544],"category_scores_gemma":[0.00002426589,0.000129548,0.00005674234,0.0001239834,0.0004183806,0.002131694,0.0001698773,0.0000995393,0.000003027541],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004330735,"about_ca_system_score_gemma":0.0003132991,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001345165,"about_ca_topic_score_gemma":0.0001391256,"domain_scores_codex":[0.9990616,0.00004456397,0.0001205239,0.0003074982,0.0002787305,0.0001871286],"domain_scores_gemma":[0.9984679,0.00004040796,0.0005117774,0.0006179665,0.0002917667,0.00007018715],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0002495395,0.0001701534,0.01224372,0.0003326342,0.0003689359,0.0003429257,0.01245692,0.01445024,0.4562185,0.4341877,0.004728445,0.06425025],"study_design_scores_gemma":[0.001638073,0.0001573245,0.004664051,0.0002247989,0.00004905336,0.00002474241,0.0002955532,0.9799593,0.01173487,0.0005575735,0.0003660462,0.0003285797],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3188851,0.00003031918,0.6765987,0.00143993,0.00005207488,0.00008719364,0.00001482328,0.00002826102,0.002863652],"genre_scores_gemma":[0.9552149,0.00001007513,0.04385437,0.0001522504,0.00002382091,2.529616e-7,0.000002897919,0.000004765018,0.0007367057],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9655091,"threshold_uncertainty_score":0.5282816,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06863170625603472,"score_gpt":0.239950264465594,"score_spread":0.1713185582095593,"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."}}