{"id":"W3110846353","doi":"10.1609/aaai.v35i16.17680","title":"Reinforced Multi-Teacher Selection for Knowledge Distillation","year":2021,"lang":"en","type":"article","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","topic":"Topic Modeling","field":"Computer Science","cited_by":121,"is_retracted":false,"has_abstract":true,"ca_institutions":"Simon Fraser University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Bottleneck; Distillation; Computer science; Artificial intelligence; Inference; Machine learning; Selection (genetic algorithm); Learning cycle; Natural language processing; Mathematics education; Mathematics","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001529737,0.001509718,0.001345738,0.0007401539,0.0007453372,0.0008011486,0.002488753,0.001571031,0.002944799],"category_scores_gemma":[0.005687374,0.0007635451,0.0008005725,0.0008743532,0.0009602632,0.002500852,0.002013067,0.002502649,0.001244233],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008102839,"about_ca_system_score_gemma":0.001584682,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003363814,"about_ca_topic_score_gemma":0.007683135,"domain_scores_codex":[0.9991096,0.0003492881,0.00003978999,0.0002383225,0.0001478297,0.0001151745],"domain_scores_gemma":[0.9980246,0.001189361,0.0001168744,0.0003154099,0.0002441794,0.000109691],"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":[0.0006267694,0.0003627529,0.001879027,0.0002137235,0.0001218693,0.0002624163,0.0003420143,0.452557,0.01457603,0.01197098,0.007217062,0.5098704],"study_design_scores_gemma":[0.00002697883,0.00004439475,0.00009275781,0.000006319121,0.00001176066,0.00002520927,0.00001506797,0.9910313,0.003254833,0.004709762,0.0007734829,0.00000821286],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04144417,0.0005857038,0.9503056,0.0004072437,0.00008085468,0.0001236943,0.000139091,0.004950334,0.001963233],"genre_scores_gemma":[0.6683879,0.0002722759,0.3239917,0.0005194155,0.0001299748,0.0003853487,0.0006924685,0.0005941559,0.00502689],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.003363814,"threshold_uncertainty_score":0.009851336,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1365034617350052,"score_gpt":0.3305311318721846,"score_spread":0.1940276701371794,"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."}}