{"id":"W2907361740","doi":"","title":"Quantization Error as a Metric for Dynamic Precision Scaling in Neural Net Training","year":2018,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"MNIST database; Rounding; Quantization (signal processing); Computer science; Scaling; Artificial neural network; Algorithm; Round-off error; Floating point; Metric (unit); Artificial intelligence; Mathematics","routes":{"ca_aff":true,"ca_fund":false,"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"],"consensus_categories":[],"category_scores_codex":[0.0002947568,0.0002810417,0.000321221,0.0006725669,0.0002111097,0.0000989091,0.00155538,0.000238464,0.00000550459],"category_scores_gemma":[0.0001447145,0.0003395853,0.0001525861,0.002124635,0.00008255665,0.0005039489,0.001023868,0.000384383,0.00002602686],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002672762,"about_ca_system_score_gemma":0.0001252391,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003143237,"about_ca_topic_score_gemma":0.00008015995,"domain_scores_codex":[0.997786,0.000109032,0.0003034711,0.001281663,0.0001019642,0.0004178017],"domain_scores_gemma":[0.9980579,0.0003909663,0.0003423567,0.0009234097,0.0001677653,0.0001176041],"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":[0.00003921518,0.00004707507,0.0003274361,0.00002947503,0.0000137805,0.00001906661,0.0003187878,0.9584336,0.00005786217,0.02833285,0.0000356686,0.01234516],"study_design_scores_gemma":[0.0003990851,0.00005978242,0.0009670184,0.00007784854,0.00002157636,0.000003925794,0.00005113746,0.9231713,0.00003492787,0.07476428,0.0001429734,0.0003061416],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2854263,0.00005230108,0.7130902,0.00009638877,0.0003406068,0.0006827083,0.000007777711,0.0001739574,0.0001297798],"genre_scores_gemma":[0.9726701,0.00006093454,0.02682879,0.00008122788,0.00006585041,0.00001055108,0.00004523975,0.00002657791,0.0002107639],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6872438,"threshold_uncertainty_score":0.9999056,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1236027287418837,"score_gpt":0.2618221371763817,"score_spread":0.138219408434498,"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."}}