{"id":"W4404672046","doi":"10.1007/978-3-031-73024-5_1","title":"How to Train the Teacher Model for Effective Knowledge Distillation","year":2024,"lang":"en","type":"book-chapter","venue":"Lecture notes in computer science","topic":"Technology-Enhanced Education Studies","field":"Social Sciences","cited_by":6,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Waterloo","funders":"","keywords":"Computer science; Distillation; Artificial intelligence","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":[0.001445114,0.0004096208,0.0003940633,0.0004793223,0.0006035963,0.002274578,0.0009349359,0.001487686,0.01744746],"category_scores_gemma":[0.00667332,0.0003521011,0.0004649448,0.0004264932,0.0007826153,0.006876411,0.001969022,0.002241718,0.005467508],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000825258,"about_ca_system_score_gemma":0.002615477,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003262254,"about_ca_topic_score_gemma":0.005308158,"domain_scores_codex":[0.9993344,0.0002856585,0.00003529382,0.0001032713,0.000186112,0.0000553438],"domain_scores_gemma":[0.9976268,0.001169194,0.0001034147,0.0003893824,0.0005050153,0.0002061502],"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":[0.0002102304,0.0007149135,0.003680252,0.0005846397,0.00006998419,0.0002479979,0.002431376,0.02419363,0.01085719,0.1860839,0.03726749,0.7336583],"study_design_scores_gemma":[0.0002106856,0.0004854055,0.002289674,0.0004802581,0.0001264082,0.0005914546,0.003421483,0.3416381,0.03383743,0.4307784,0.1860477,0.00009297264],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03395286,0.0006724981,0.8714542,0.01568909,0.0003470705,0.000375949,0.0003018225,0.002875966,0.07433061],"genre_scores_gemma":[0.3437329,0.0007821783,0.6040977,0.001223412,0.00007549599,0.0005007227,0.0004185886,0.0003855157,0.04878359],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01744746,"threshold_uncertainty_score":0.05836761,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03277099080217485,"score_gpt":0.338730442324303,"score_spread":0.3059594515221282,"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."}}