{"id":"W3211782300","doi":"10.48550/arxiv.2012.15495","title":"Towards Zero-Shot Knowledge Distillation for Natural Language Processing","year":2020,"lang":"en","type":"article","venue":"arXiv (Cornell University)","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Carleton University","funders":"","keywords":"Computer science; Distillation; Task (project management); Benchmark (surveying); Artificial intelligence; Knowledge transfer; Natural language processing; Variety (cybernetics); Domain knowledge; Machine learning; Transfer of learning; Domain (mathematical analysis); Shot (pellet); Knowledge management; Mathematics; Engineering","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":[0.002442189,0.001341886,0.001128529,0.0009144751,0.0008427227,0.001686172,0.002322563,0.002048057,0.003253731],"category_scores_gemma":[0.008751475,0.0005392545,0.0009627036,0.0009587877,0.002047176,0.005159403,0.004749048,0.004868771,0.001733781],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001151102,"about_ca_system_score_gemma":0.001750164,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003486001,"about_ca_topic_score_gemma":0.005690702,"domain_scores_codex":[0.9985514,0.0005511214,0.00005756849,0.0003383349,0.000367529,0.0001339445],"domain_scores_gemma":[0.9967306,0.002089448,0.000123077,0.0006708782,0.0002521943,0.0001338448],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0007581346,0.0004982119,0.001485825,0.0006650536,0.0001597548,0.000307487,0.0006650162,0.4507914,0.01354203,0.09692349,0.02161038,0.4125933],"study_design_scores_gemma":[0.00002863878,0.00006382809,0.00008908231,0.00002300302,0.00001026769,0.00004623795,0.00004617133,0.9294652,0.00481591,0.06304222,0.00235815,0.00001129073],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.05212231,0.001332996,0.932523,0.001596081,0.000142315,0.0001135125,0.0005897223,0.006052371,0.005527645],"genre_scores_gemma":[0.5955835,0.0007924551,0.3899967,0.0009699405,0.0002159253,0.0002709607,0.002768278,0.0008521605,0.008549999],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.003486001,"threshold_uncertainty_score":0.01291573,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.09527768866442518,"score_gpt":0.21964052195364,"score_spread":0.1243628332892149,"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."}}