{"id":"W2903735800","doi":"10.1109/fpl.2018.00014","title":"Embracing Diversity: Enhanced DSP Blocks for Low-Precision Deep Learning on FPGAs","year":2018,"lang":"en","type":"article","venue":"","topic":"Low-power high-performance VLSI design","field":"Engineering","cited_by":67,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Digital signal processing; Computer science; Field-programmable gate array; Convolutional neural network; Floating point; Embedded system; Block (permutation group theory); Parallel computing; Overhead (engineering); Computer hardware; Algorithm; Artificial intelligence; Mathematics","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003409242,0.0008721348,0.0002706492,0.0006319354,0.0002154651,0.0006463884,0.001450719,0.0003263655,0.01127411],"category_scores_gemma":[0.0009412182,0.0003064674,0.0002932677,0.000573071,0.0002392887,0.001174949,0.0005572658,0.0008207788,0.002550297],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005197991,"about_ca_system_score_gemma":0.0006585142,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002256708,"about_ca_topic_score_gemma":0.003861826,"domain_scores_codex":[0.9996821,0.00003766244,0.00002217983,0.00006266813,0.0001211112,0.0000744179],"domain_scores_gemma":[0.9996794,0.0001013449,0.00003867203,0.00006789632,0.00008053159,0.00003197929],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.002004507,0.0004323852,0.005433514,0.0007297008,0.000211265,0.0007549066,0.0002080456,0.1043631,0.1748856,0.0215491,0.03820482,0.6512229],"study_design_scores_gemma":[0.0006363843,0.002124311,0.003841812,0.0001761114,0.0001773466,0.0007663138,0.0001392007,0.652987,0.2349801,0.008746142,0.09532485,0.0001004795],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3273567,0.003328562,0.6032562,0.000675976,0.0006310618,0.0003807598,0.001189786,0.02391854,0.03926243],"genre_scores_gemma":[0.7708448,0.000503013,0.2168541,0.0004663967,0.00007822984,0.0002064763,0.001061638,0.0003530485,0.009632283],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01127411,"threshold_uncertainty_score":0.03771561,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.00865661826823669,"score_gpt":0.2133980077770524,"score_spread":0.2047413895088157,"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."}}