{"id":"W2552737632","doi":"10.1038/ncomms13276","title":"Random synaptic feedback weights support error backpropagation for deep learning","year":2016,"lang":"en","type":"article","venue":"Nature Communications","topic":"Advanced Memory and Neural Computing","field":"Engineering","cited_by":851,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University; University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Vetenskapsrådet","keywords":"Backpropagation; Computer science; Mechanism (biology); Constraint (computer-aided design); Mistake; Artificial intelligence; Blame; Artificial neural network; Synaptic weight; Mathematics","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":[0.0001628266,0.0001214164,0.0001393317,0.00006044486,0.0003119153,0.00001602739,0.0004408282,0.0001546021,0.00002613653],"category_scores_gemma":[0.000241586,0.00009193808,0.00007414469,0.0001318665,0.00004952373,0.0001936676,0.00007640135,0.0004650945,0.000053045],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005736767,"about_ca_system_score_gemma":0.000009546193,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":1.974682e-7,"about_ca_topic_score_gemma":0.00002151855,"domain_scores_codex":[0.9993472,0.00005425928,0.0002002035,0.000129003,0.00007394598,0.0001953569],"domain_scores_gemma":[0.9983657,0.0008216085,0.00005227635,0.0006150919,0.00009340319,0.00005192025],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0003990516,0.0002248958,0.002323291,0.0003929506,0.0005796215,0.000006322924,0.001846796,0.03437373,0.4452102,0.05117656,0.006237568,0.457229],"study_design_scores_gemma":[0.01000796,0.0003433772,0.00502016,0.0005366371,0.0002577355,0.00008573907,0.0003358149,0.2427692,0.09100296,0.009359598,0.6384996,0.001781328],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3056303,0.01331144,0.649169,0.007242402,0.00216674,0.002692105,0.00005054013,0.003468198,0.01626931],"genre_scores_gemma":[0.983717,0.0003167998,0.01536029,0.00006281456,0.00008310643,0.00006512364,0.00004697995,0.00003283903,0.00031509],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6780866,"threshold_uncertainty_score":0.3749126,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01933405323896934,"score_gpt":0.2786935998621625,"score_spread":0.2593595466231931,"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."}}