{"id":"W4230481654","doi":"10.1109/aspdac.2018.8297303","title":"Quad-multiplier packing based on customized floating point for convolutional neural networks on FPGA","year":2018,"lang":"en","type":"article","venue":"2018 23rd Asia and South Pacific Design Automation Conference (ASP-DAC)","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Waseda University; Canadian Institute for Advanced Research","keywords":"Field-programmable gate array; Floating point; Computer science; Convolutional neural network; Adder; Multiplier (economics); Throughput; Computation; Artificial neural network; Deep learning; Reduction (mathematics); Parallel computing; Computer hardware; Inference; Computer engineering; Digital signal processing; Artificial intelligence; Algorithm; Latency (audio)","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0007987877,0.0004455405,0.0004236521,0.0002344329,0.0009438337,0.0004125922,0.000578953,0.0002137508,0.00007797659],"category_scores_gemma":[0.0002618969,0.0004113959,0.0001279255,0.0004719076,0.0002743547,0.0005494349,0.00009159264,0.0003021173,0.0001359006],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000112811,"about_ca_system_score_gemma":0.0001531739,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003268913,"about_ca_topic_score_gemma":0.00000217447,"domain_scores_codex":[0.9968566,0.0003369458,0.0006360934,0.0009985799,0.0004856165,0.0006861979],"domain_scores_gemma":[0.9968166,0.001223289,0.0005050774,0.0007792703,0.0004259585,0.0002498348],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002094041,0.0007033449,0.001461375,0.0001464475,0.0001850462,0.00001732343,0.01056962,0.3060441,0.006420207,0.275167,0.01823223,0.3789592],"study_design_scores_gemma":[0.001806308,0.0004176039,0.001173471,0.0001063489,0.00001684644,0.000005773397,0.0002779443,0.9913988,0.0009099584,0.002755681,0.0006673264,0.00046398],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002648756,0.00001385867,0.9915525,0.001260148,0.0005976444,0.001454951,0.00002034335,0.0006050955,0.001846731],"genre_scores_gemma":[0.8945667,0.00000323462,0.1039316,0.000572359,0.0003658992,0.0002644385,0.00005057202,0.00003497029,0.0002102948],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8919179,"threshold_uncertainty_score":0.9998338,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04641943679345872,"score_gpt":0.2727081740530657,"score_spread":0.226288737259607,"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."}}