{"id":"W3196726140","doi":"10.1109/isit45174.2021.9518254","title":"Differentially Quantized Gradient Descent","year":2021,"lang":"en","type":"article","venue":"","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":14,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Russian Academy of Sciences; Natural Sciences and Engineering Research Council of Canada; Princeton University; Jet Propulsion Laboratory; Division of Mathematical Sciences; Moscow Institute of Physics and Technology; King Abdullah University of Science and Technology; University of Ottawa; Indian Institute of Science; University of Tehran; National Aeronautics and Space Administration; California Institute of Technology; Amgen; National Science Foundation","keywords":"Gradient descent; Stochastic gradient descent; Quantization (signal processing); Convergence (economics); Descent (aeronautics); Computer science; Algorithm; Mathematics; Discrete mathematics; Theoretical computer science; Combinatorics; Artificial intelligence; Physics","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0000801963,0.0001037484,0.0001247854,0.00007206167,0.00007020156,0.0001545865,0.0004485155,0.00003481664,0.0001500331],"category_scores_gemma":[0.00006865621,0.00009245633,0.00006864666,0.0003716036,0.00002091113,0.0001544584,0.0002768303,0.00005818433,0.00004385854],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003438873,"about_ca_system_score_gemma":0.00006496326,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000005739304,"about_ca_topic_score_gemma":0.000005542449,"domain_scores_codex":[0.998969,0.00004352735,0.0001915272,0.0003408582,0.0002405853,0.0002145379],"domain_scores_gemma":[0.9991148,0.00004395482,0.00004422504,0.0005439279,0.0001587885,0.0000942992],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000001491676,0.0001204911,0.0001926921,0.000004458424,0.00001205511,0.00002385269,0.0001184662,0.0000484271,0.002133591,0.9901888,0.001569785,0.00558591],"study_design_scores_gemma":[0.00205494,0.0002474148,0.00888725,0.00009229683,0.00003355111,0.0002300208,0.00007533779,0.5144119,0.3710314,0.09633248,0.005471998,0.00113138],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.00256527,0.00005247429,0.9901201,0.0009205802,0.0004346019,0.0001005233,4.832936e-7,0.0006751452,0.005130791],"genre_scores_gemma":[0.4102889,0.00002149983,0.5880877,0.0004728695,0.00001926499,0.00001698929,0.000004814108,0.000007737406,0.001080288],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.8938563,"threshold_uncertainty_score":0.377026,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01959072728710984,"score_gpt":0.2404384011253772,"score_spread":0.2208476738382674,"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."}}