{"id":"W2894812324","doi":"10.1016/j.disopt.2023.100795","title":"Principled deep neural network training through linear programming","year":2023,"lang":"en","type":"article","venue":"Discrete Optimization","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":25,"is_retracted":false,"has_abstract":false,"ca_institutions":"","funders":"Office of Naval Research; Institut de Valorisation des Données; National Science Foundation","keywords":"Computer science; Deep learning; Polyhedron; Artificial intelligence; Artificial neural network; Linear programming; Perspective (graphical); Dependency (UML); Task (project management); Representation (politics); Sample (material); Function (biology); Machine learning; Mathematical optimization; Algorithm; Theoretical computer science; Mathematics; Geometry","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":[0.001450398,0.0009229489,0.0009681083,0.0003809315,0.0003972864,0.001003368,0.001503219,0.001423128,0.004876105],"category_scores_gemma":[0.003782729,0.001081147,0.0005469181,0.0005608174,0.001183353,0.00141484,0.002071309,0.002833828,0.0009597789],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000806882,"about_ca_system_score_gemma":0.001462002,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001873661,"about_ca_topic_score_gemma":0.003124789,"domain_scores_codex":[0.9995056,0.0001923573,0.00001979916,0.00008117711,0.0001662712,0.00003480553],"domain_scores_gemma":[0.9989153,0.0007404126,0.00006729219,0.00008408965,0.0001451695,0.00004781233],"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":[0.00006694903,0.00005718154,0.0001845737,0.0001341213,0.00003966111,0.00002795144,0.00004516065,0.8533142,0.001639038,0.06917565,0.003935243,0.07138035],"study_design_scores_gemma":[0.00000374427,0.000005984513,0.00000904125,0.000003761801,0.000001558305,0.000003683279,0.000001097775,0.9927359,0.0002096277,0.006788405,0.0002359003,0.000001321426],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001525642,0.00006342278,0.9969339,0.0001378653,0.00002184199,0.00001727088,0.00002171087,0.0001868987,0.001091348],"genre_scores_gemma":[0.2062439,0.0002811968,0.7820092,0.0003607417,0.0001175602,0.0004440136,0.00018523,0.0005856634,0.009772485],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.004876105,"threshold_uncertainty_score":0.01631218,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04017734624734458,"score_gpt":0.2895398856964629,"score_spread":0.2493625394491183,"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."}}