{"id":"W2963375484","doi":"","title":"PAC-Bayes bounds for stable algorithms with instance-dependent priors","year":2018,"lang":"en","type":"article","venue":"HAL (Le Centre pour la Communication Scientifique Directe)","topic":"Machine Learning and Algorithms","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Defence Science and Technology Group; DeepMind; Natural Sciences and Engineering Research Council of Canada; National Natural Science Foundation of China; Simons Institute for the Theory of Computing, University of California Berkeley; Office of Defense Programs; Multidisciplinary University Research Initiative; Engineering and Physical Sciences Research Council; Defence Science and Technology Laboratory; U.S. Department of Defense; Alberta Innovates; Alberta Machine Intelligence Institute","keywords":"Prior probability; Bayes' theorem; Stability (learning theory); Algorithm; Mathematics; Gaussian; Artificial intelligence; Computer science; Upper and lower bounds; Naive Bayes classifier; Classifier (UML); Pattern recognition (psychology); Bayesian probability; Support vector machine; Machine learning","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02255722,0.003487689,0.002697654,0.003761004,0.002149279,0.005700478,0.004187939,0.003994776,0.007489618],"category_scores_gemma":[0.1566048,0.001540666,0.002666752,0.002699687,0.006525961,0.01183442,0.006139093,0.01084156,0.002185066],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004935547,"about_ca_system_score_gemma":0.003493574,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001503366,"about_ca_topic_score_gemma":0.001581557,"domain_scores_codex":[0.9797773,0.00778232,0.0009694032,0.002893615,0.00727006,0.001307349],"domain_scores_gemma":[0.8428707,0.1319953,0.005092165,0.009169442,0.009091675,0.00178075],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0004226025,0.000114288,0.001549533,0.0005113074,0.0002008974,0.0001623465,0.0003585706,0.2483013,0.002782222,0.6655042,0.003985277,0.07610747],"study_design_scores_gemma":[0.00001820141,0.0001087951,0.0003204603,0.0001365061,0.0000494128,0.000118595,0.00003293128,0.6833974,0.002842539,0.3107728,0.002157487,0.00004496221],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003602884,0.001650912,0.9900102,0.000445269,0.00009797851,0.00005797334,0.0000621994,0.0002429883,0.003829552],"genre_scores_gemma":[0.3434656,0.00389498,0.6386814,0.001477448,0.001127272,0.001189543,0.0007007752,0.001731244,0.00773166],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02255722,"threshold_uncertainty_score":0.1192954,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.00848001380485293,"score_gpt":0.2249546404959976,"score_spread":0.2164746266911447,"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."}}