{"id":"W3171596153","doi":"10.1109/jsac.2021.3087272","title":"LOSP: Overlap Synchronization Parallel With Local Compensation for Fast Distributed Training","year":2021,"lang":"en","type":"article","venue":"IEEE Journal on Selected Areas in Communications","topic":"Stochastic Gradient Optimization Techniques","field":"Computer Science","cited_by":21,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"National Key Research and Development Program of China; Research Grants Council, University Grants Committee; China Postdoctoral Science Foundation; Impact Fund; Science, Technology and Innovation Commission of Shenzhen Municipality; National Natural Science Foundation of China","keywords":"Computer science; Scalability; Synchronization (alternating current); Computation; Speedup; Distributed computing; Overhead (engineering); Convergence (economics); Stochastic gradient descent; Rate of convergence; Compensation (psychology); Data synchronization; Mathematical optimization; Parallel computing; Algorithm; Key (lock); Computer network; Artificial intelligence; Artificial neural network","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":[],"consensus_categories":[],"category_scores_codex":[0.0003112482,0.000173023,0.0002340419,0.0002478045,0.0004662334,0.0002590248,0.001084053,0.00008993105,0.000007930944],"category_scores_gemma":[0.0003214473,0.0001678626,0.00005290133,0.001690856,0.00009852886,0.000447622,0.00008461929,0.0004569,0.000003302319],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004067324,"about_ca_system_score_gemma":0.0005720269,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00000458572,"about_ca_topic_score_gemma":0.00008633645,"domain_scores_codex":[0.9983737,0.0002635555,0.0004855109,0.000266795,0.000320174,0.000290289],"domain_scores_gemma":[0.996714,0.0006320817,0.0003110323,0.0009732004,0.001257835,0.0001118681],"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":[0.0001654565,0.001529207,0.002775835,0.00003165619,0.0002115296,0.00004130507,0.00364333,0.6034523,0.0006457759,0.3254565,0.002215389,0.05983171],"study_design_scores_gemma":[0.001540602,0.0003153134,0.00333842,0.0003063427,0.00002416334,0.0003910182,0.0002974968,0.9854295,0.0007112382,0.006544999,0.0007999973,0.000300925],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.001079609,0.0000879177,0.9952255,0.002631913,0.0001225191,0.0003316126,0.00002048382,0.0001789216,0.0003214924],"genre_scores_gemma":[0.6904431,0.00008722878,0.3089691,0.0001832148,0.00003111215,0.00006973521,0.0001819175,0.00001627365,0.00001825231],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6893635,"threshold_uncertainty_score":0.6845239,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03917666790837083,"score_gpt":0.2856775653390298,"score_spread":0.246500897430659,"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."}}