{"id":"W1996911291","doi":"10.1016/j.neunet.2005.06.034","title":"Improving dimensionality reduction with spectral gradient descent","year":2005,"lang":"en","type":"article","venue":"Neural Networks","topic":"Face and Expression Recognition","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Toronto","funders":"Canada Research Chairs; Government of Canada; Canadian Institute for Advanced Research","keywords":"Gradient descent; Dimensionality reduction; Eigenvalues and eigenvectors; Maxima and minima; Curse of dimensionality; Matrix (chemical analysis); Reduction (mathematics); Computer science; Eigendecomposition of a matrix; Mathematics; Mathematical optimization; Gradient method; Iterative method; Algorithm; Artificial neural network; Artificial intelligence","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00009594207,0.000109977,0.00008777811,0.00003182275,0.0001635969,0.00008449521,0.0001844022,0.00004561825,0.0000130704],"category_scores_gemma":[0.000002916092,0.00008142469,0.00004036711,0.0001682706,0.00002757508,0.0004899537,0.00007324528,0.0001868269,0.00001589054],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004670253,"about_ca_system_score_gemma":0.00001126035,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001852898,"about_ca_topic_score_gemma":0.000009979512,"domain_scores_codex":[0.9990376,0.00004297565,0.0001368598,0.0003116743,0.0001929597,0.0002779877],"domain_scores_gemma":[0.9995355,0.0000157601,0.00006736346,0.0002391845,0.0000428248,0.00009934587],"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":[0.0001008407,0.0002031314,0.0004576497,0.00001002659,0.00001525739,0.0000248015,0.0002042469,0.1779641,0.006151091,0.002341831,0.00696883,0.8055582],"study_design_scores_gemma":[0.0002892271,0.0001131864,0.002004393,0.00002481735,0.000005606955,0.00007977448,0.000009556711,0.9925218,0.00390169,0.0001991982,0.0006926232,0.0001581183],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7407039,0.0001792663,0.2549582,0.002553004,0.0007716521,0.0001973829,4.803906e-7,0.0002560037,0.0003801263],"genre_scores_gemma":[0.9836342,0.00001117766,0.01537937,0.0004650207,0.0004176884,0.000009003511,0.000005239431,0.000006347561,0.00007196588],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8145577,"threshold_uncertainty_score":0.3320403,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01126066845199177,"score_gpt":0.2120135920215758,"score_spread":0.200752923569584,"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."}}