{"id":"W4399511705","doi":"10.48550/arxiv.2406.04512","title":"To Distill or Not to Distill? On the Robustness of Robust Knowledge Distillation","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Bayesian Modeling and Causal Inference","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Alliance de recherche numérique du Canada; Social Sciences and Humanities Research Council of Canada; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Robustness (evolution); Distillation; Computer science; Biochemical engineering; Process engineering; Engineering; Chromatography; Chemistry","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0003716523,0.0003992092,0.0004023358,0.0003128115,0.0001519296,0.000202592,0.002265545,0.0002228322,0.00003105826],"category_scores_gemma":[0.0001728711,0.0002937287,0.0001943723,0.001289251,0.00009195667,0.0000967742,0.003179983,0.0006149929,0.0002726805],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002379068,"about_ca_system_score_gemma":0.0002948616,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008993553,"about_ca_topic_score_gemma":0.0001452286,"domain_scores_codex":[0.9977115,0.0001505326,0.0003100621,0.001292334,0.0001694225,0.0003661647],"domain_scores_gemma":[0.9973179,0.0003440999,0.0001551861,0.001609745,0.000299393,0.0002737124],"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":[0.00009335305,0.00006946464,0.00002825134,0.0001150056,0.00003728016,0.00003615761,0.0004951971,0.7449367,0.00001650907,0.2507067,0.001244982,0.002220411],"study_design_scores_gemma":[0.0001036603,0.0001478202,0.0007651107,0.0006262127,0.00006643498,0.000002140769,0.00006397082,0.9877566,0.0001931298,0.00921763,0.0005729145,0.0004843678],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1618398,0.00001552779,0.833259,0.0009959395,0.00106742,0.000432449,0.00007505604,0.0001950469,0.002119745],"genre_scores_gemma":[0.9937285,0.00001710215,0.001963063,0.0001409895,0.0001407416,0.000004279505,0.00000863736,0.00002691412,0.003969753],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8318887,"threshold_uncertainty_score":0.9999515,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1658194347244474,"score_gpt":0.2395218233615591,"score_spread":0.07370238863711168,"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."}}