{"id":"W2971946864","doi":"10.1109/tc.2019.2936192","title":"New Flexible Multiple-Precision Multiply-Accumulate Unit for Deep Neural Network Training and Inference","year":2019,"lang":"en","type":"article","venue":"IEEE Transactions on Computers","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":55,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Saskatchewan","funders":"Natural Sciences and Engineering Research Council of Canada; University of Saskatchewan","keywords":"Floating-point unit; Computer science; Floating point; IEEE floating point; Binary number; Unit (ring theory); Artificial neural network; 16-bit; Single-precision floating-point format; Algorithm; Computer hardware; Overhead (engineering); Exponent; Arithmetic; Control unit; Parallel computing; Mathematics; Artificial intelligence","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0001312148,0.0002939641,0.0003006039,0.0001456115,0.0003660649,0.0001799607,0.0007314142,0.00009937916,0.00001045098],"category_scores_gemma":[0.000006775736,0.0003020143,0.0001234043,0.0006328118,0.00004238851,0.0006853032,0.00001748393,0.0002988908,0.00003842933],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004304483,"about_ca_system_score_gemma":0.00005131188,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001998776,"about_ca_topic_score_gemma":0.00003335709,"domain_scores_codex":[0.998109,0.00005686476,0.0003476756,0.0007431018,0.0002130585,0.0005302517],"domain_scores_gemma":[0.9970868,0.001770568,0.0001211755,0.0006867619,0.0000744654,0.0002601878],"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":[0.00002726454,0.00001829259,0.00003823456,0.000004803059,0.00001273602,6.278415e-7,0.0002263541,0.6157026,0.0002696801,0.0004688373,0.00005370522,0.3831768],"study_design_scores_gemma":[0.001304941,0.0002368727,0.0003343974,0.00005207208,0.00001137076,0.00001379032,0.00001617815,0.993104,0.0009420931,0.00145479,0.002201666,0.0003277777],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01320452,0.00004725929,0.9832313,0.0004406256,0.001607877,0.0009581172,0.000006284915,0.0004678474,0.00003616955],"genre_scores_gemma":[0.6293538,0.0000186615,0.3699071,0.000413451,0.00009087643,0.00004920349,0.000002862153,0.00002271529,0.0001412891],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.6161493,"threshold_uncertainty_score":0.9999432,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05346103363208157,"score_gpt":0.3028072657203022,"score_spread":0.2493462320882206,"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."}}