{"id":"W3213992252","doi":"10.1109/istc49272.2021.9594251","title":"Exploitation of temporal structure in momentum-SGD for gradient compression","year":2021,"lang":"en","type":"article","venue":"","topic":"Advanced Data Compression Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"Science and Engineering Research Council; Huawei Technologies","keywords":"Bottleneck; Leverage (statistics); Computer science; Momentum (technical analysis); Correlation; Volume (thermodynamics); Computer engineering; Artificial intelligence; Mathematics; Embedded system","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":[0.00007734836,0.00008435851,0.0001494051,0.00009980712,0.00003020348,0.00002294183,0.0003882463,0.00004120021,0.0000220112],"category_scores_gemma":[0.00004780083,0.00007126964,0.00003396346,0.0002680076,0.00001669265,0.0005200501,0.0002776849,0.00006154628,4.63438e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002955887,"about_ca_system_score_gemma":0.00003514907,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001161944,"about_ca_topic_score_gemma":0.00001704302,"domain_scores_codex":[0.999097,0.00004139687,0.0002681286,0.0002845267,0.0001745166,0.0001343547],"domain_scores_gemma":[0.9991899,0.0000851946,0.0001051624,0.0004575504,0.0001254148,0.00003674636],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.00004665025,0.0003734153,0.00321439,0.0001746784,0.00001364161,0.00002079218,0.001242481,0.0009983343,0.5219787,0.3429731,0.01950923,0.1094546],"study_design_scores_gemma":[0.0004766281,0.00007203718,0.00120935,0.00008273614,0.000001253728,0.000003889328,0.0001051084,0.03011293,0.8639075,0.09836044,0.005538527,0.0001296487],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03106566,0.0000874639,0.9679104,0.0002832586,0.0001439183,0.0002187205,0.00002160139,0.0001030713,0.0001658691],"genre_scores_gemma":[0.4959479,0.000009119844,0.5037975,0.00007358785,0.000007273685,0.00002519916,0.00003691236,0.000004128987,0.00009838099],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.4648822,"threshold_uncertainty_score":0.2906292,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02139617175792598,"score_gpt":0.2981553570402896,"score_spread":0.2767591852823637,"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."}}