{"id":"W4387559563","doi":"10.48550/arxiv.2310.04935","title":"Statistical Guarantees for Variational Autoencoders using PAC-Bayesian Theory","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Generative Adversarial Networks and Image Synthesis","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université Laval","funders":"Institut de Valorisation des Données; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Generalization; Bayesian probability; Computer science; Generative grammar; Upper and lower bounds; Posterior probability; Distribution (mathematics); Artificial intelligence; Generative model; Statistical learning theory; Algorithm; Machine learning; 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":[0.008635503,0.002425255,0.001943486,0.001506177,0.001178297,0.002937565,0.003276421,0.003008878,0.004435803],"category_scores_gemma":[0.04311458,0.001660662,0.001690266,0.001423479,0.004815209,0.00674262,0.005293379,0.008854601,0.0009745539],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.003486539,"about_ca_system_score_gemma":0.003264036,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003339439,"about_ca_topic_score_gemma":0.004443485,"domain_scores_codex":[0.9957624,0.001481532,0.000198811,0.000632967,0.001619484,0.0003048131],"domain_scores_gemma":[0.9702543,0.02432909,0.001273551,0.001616019,0.001886886,0.000640168],"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":[0.00006763382,0.00004550647,0.0005492477,0.000172014,0.00007595489,0.0001020293,0.0001410441,0.3988551,0.001402753,0.577947,0.002295902,0.01834588],"study_design_scores_gemma":[0.00000955838,0.00001795334,0.00008759247,0.00003783551,0.000007184983,0.00003580473,0.00001215035,0.714845,0.0005445101,0.2836178,0.0007713767,0.00001323079],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.004273006,0.0005133604,0.9909532,0.0007173171,0.00003386006,0.00002368622,0.000103921,0.0001701486,0.003211444],"genre_scores_gemma":[0.5970052,0.004676335,0.3826107,0.00178245,0.0006333332,0.0007040714,0.001173435,0.001160637,0.01025374],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.008635503,"threshold_uncertainty_score":0.0456695,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1127911845074769,"score_gpt":0.2227502722686008,"score_spread":0.1099590877611239,"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."}}