{"id":"W7039154034","doi":"","title":"Limitations of information : theoretic generalization bounds for gradient descent methods in stochastic convex optimization","year":2023,"lang":"en","type":"article","venue":"KTH Publication Database DiVA (KTH Royal Institute of Technology)","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Vetenskapsrådet; University of Toronto","keywords":"Minimax; Generalization; Stochastic gradient descent; Gradient descent; Mutual information; Regular polygon; Convex optimization; Convex function","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":[0.05462456,0.002990184,0.004158111,0.003283205,0.00220929,0.007251472,0.007455405,0.004710798,0.00540819],"category_scores_gemma":[0.218569,0.001734627,0.002906425,0.003080671,0.01177481,0.02158273,0.009168388,0.017717,0.001596908],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005450668,"about_ca_system_score_gemma":0.003789011,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002821594,"about_ca_topic_score_gemma":0.001909511,"domain_scores_codex":[0.9598382,0.0198341,0.002151997,0.00338631,0.0136239,0.001165458],"domain_scores_gemma":[0.7539932,0.2090331,0.005897414,0.01726553,0.01237382,0.001436927],"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":[0.0001792764,0.00006607267,0.000568704,0.0005643258,0.0001453651,0.00009097026,0.0003139152,0.09090061,0.0005277299,0.8617806,0.003644835,0.04121764],"study_design_scores_gemma":[0.00002533405,0.0000999227,0.0002277396,0.0003411428,0.00003717307,0.0001039518,0.00006936265,0.3012605,0.001080186,0.6927329,0.003971112,0.00005064319],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.006454091,0.009499057,0.9570366,0.008339345,0.0003715137,0.000105283,0.0002337121,0.0003433135,0.01761703],"genre_scores_gemma":[0.6050841,0.01612291,0.3606079,0.005254993,0.002452809,0.001130671,0.0005607734,0.001254444,0.007531437],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.05462456,"threshold_uncertainty_score":0.2888857,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06181125478238244,"score_gpt":0.3177407573990477,"score_spread":0.2559295026166653,"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."}}