{"id":"W3034007088","doi":"10.3389/fnins.2021.633674","title":"Scaling Equilibrium Propagation to Deep ConvNets by Drastically Reducing Its Gradient Estimator Bias","year":2021,"lang":"en","type":"preprint","venue":"Frontiers in Neuroscience","topic":"Domain Adaptation and Few-Shot Learning","field":"Computer Science","cited_by":10,"is_retracted":false,"has_abstract":true,"ca_institutions":"Canadian Institute for Advanced Research; Université de Montréal","funders":"Samsung; European Research Council; Natural Sciences and Engineering Research Council of Canada; Canadian Institute for Advanced Research","keywords":"Estimator; MNIST database; Computer science; Artificial neural network; Scaling; Backpropagation; Cross entropy; Deep learning; Algorithm; Artificial intelligence; Applied mathematics; Mathematics; Pattern recognition (psychology); Statistics","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":["metaepi_narrow","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.001010673,0.0004001314,0.0004806082,0.0004930988,0.0002263944,0.001490158,0.00218997,0.0001815841,0.00000525405],"category_scores_gemma":[0.003188329,0.000443458,0.0001051224,0.001448743,0.0001663545,0.0007459266,0.001991285,0.0008626286,0.00001255777],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002687667,"about_ca_system_score_gemma":0.0004958373,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001601999,"about_ca_topic_score_gemma":0.000001376771,"domain_scores_codex":[0.9951242,0.0003533384,0.0007145156,0.001994504,0.001003989,0.0008094331],"domain_scores_gemma":[0.9980158,0.0001339307,0.000331296,0.0008647733,0.000198934,0.0004552511],"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":[0.00003293292,0.0004080638,0.003052375,0.0003990522,0.00001366041,0.0004585923,0.007683922,0.777631,0.1314124,0.003635515,0.007392231,0.06788022],"study_design_scores_gemma":[0.0001850116,0.00006323647,0.001923726,0.0003100724,0.000005964882,0.00001949946,0.0001253416,0.9904531,0.004831356,0.0002342902,0.001364646,0.0004838095],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0593147,0.0003360461,0.9287459,0.001274941,0.009026481,0.0006389421,0.000005400229,0.0002240668,0.0004334834],"genre_scores_gemma":[0.723728,0.00003356422,0.2746351,0.001191898,0.00006012049,0.00008291767,0.00001349665,0.00003236132,0.0002224972],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6644133,"threshold_uncertainty_score":0.9998017,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03290051895411088,"score_gpt":0.2678406362161545,"score_spread":0.2349401172620436,"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."}}