{"id":"W2117499988","doi":"","title":"The Difficulty of Training Deep Architectures and the Effect of Unsupervised Pre-Training","year":2009,"lang":"en","type":"article","venue":"International Conference on Artificial Intelligence and Statistics","topic":"Generative Adversarial Networks and Image Synthesis","field":"Computer Science","cited_by":323,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal","funders":"","keywords":"Artificial intelligence; Computer science; Initialization; Robustness (evolution); Deep learning; Training (meteorology); Machine learning; Deep neural networks; Unsupervised learning; Training set","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005936504,0.0001189173,0.0001958824,0.00004166448,0.0001689681,0.0001566529,0.0004354919,0.00002891906,0.000009269998],"category_scores_gemma":[0.0007153387,0.00006370814,0.00003697781,0.00008065949,0.000459485,0.00005390874,0.00005569533,0.0001179408,6.481765e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000005448499,"about_ca_system_score_gemma":0.0000284661,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002594124,"about_ca_topic_score_gemma":0.00005611958,"domain_scores_codex":[0.998877,0.0002149644,0.0003309037,0.0001847301,0.0002554294,0.0001369002],"domain_scores_gemma":[0.9974318,0.002075209,0.0001520812,0.0001508773,0.0001524195,0.00003765902],"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":[0.0001662374,0.00000712588,0.00001116603,0.000002212583,0.00001933109,8.856098e-7,0.003680728,0.001372304,0.0008520117,0.4389897,0.000004809095,0.5548935],"study_design_scores_gemma":[0.0001121132,0.0004572345,0.001142773,0.00004265944,0.00001288791,0.000004723776,0.0006535111,0.80071,0.008951396,0.1878019,0.00002558819,0.00008518693],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02852775,0.0000779141,0.9686359,0.00160158,0.0001937476,0.000160776,0.0000205707,0.000009007502,0.0007727653],"genre_scores_gemma":[0.9907714,0.0001286131,0.008945464,0.00007481471,0.00005237582,0.000005289258,0.000002455883,0.000002542735,0.00001703877],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9622437,"threshold_uncertainty_score":0.2597942,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05197575815289519,"score_gpt":0.3078173894272081,"score_spread":0.2558416312743129,"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."}}