{"id":"W3035838478","doi":"10.1016/j.ijar.2020.02.007","title":"Discriminative training of feed-forward and recurrent sum-product networks by extended Baum-Welch","year":2020,"lang":"en","type":"article","venue":"International Journal of Approximate Reasoning","topic":"Face and Expression Recognition","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":false,"ca_institutions":"Vector Institute; University of Waterloo","funders":"","keywords":"Discriminative model; Gradient descent; Monotonic function; Robustness (evolution); Computer science; Artificial intelligence; Maximization; Artificial neural network; Product (mathematics); Generative grammar; Pattern recognition (psychology); Machine learning; Algorithm; Mathematics; Mathematical optimization","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":[0.0004186569,0.0001352978,0.0002622294,0.0001031434,0.00005128832,0.0001305604,0.0006043765,0.00004021823,0.00001255626],"category_scores_gemma":[0.0002593363,0.0001116304,0.00009551259,0.000145034,0.00005314256,0.0007469404,0.0001935235,0.0002668489,0.000001158458],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003491587,"about_ca_system_score_gemma":0.00005009925,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004755301,"about_ca_topic_score_gemma":3.348312e-7,"domain_scores_codex":[0.9985077,0.00007266155,0.0004838801,0.0002307146,0.0005330683,0.000172004],"domain_scores_gemma":[0.9986453,0.00009691725,0.0005934592,0.00008563814,0.0004287577,0.0001499009],"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":[0.0005791715,0.0003009175,0.0009556114,0.00009592833,0.0005418828,0.0001257311,0.02188821,0.001389422,0.02940588,0.01125436,0.009013642,0.9244493],"study_design_scores_gemma":[0.003835922,0.001123367,0.002808173,0.002761379,0.0001074934,0.0007313092,0.005419868,0.8856431,0.0869102,0.005627134,0.004242308,0.0007897323],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2243808,0.001906606,0.7643156,0.006882824,0.001090348,0.0001520842,0.00002043103,0.00003732592,0.001214033],"genre_scores_gemma":[0.9690883,0.0004377148,0.03003545,0.000180016,0.0002260302,0.000002567253,0.000009252336,0.000009364618,0.00001136444],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9236595,"threshold_uncertainty_score":0.4552157,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02659934367792731,"score_gpt":0.2750052373150831,"score_spread":0.2484058936371558,"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."}}