{"id":"W3046245004","doi":"10.1109/ipdpsw50202.2020.00034","title":"Optimizing OpenCL Kernels and Runtime for DNN Inference on FPGAs","year":2020,"lang":"en","type":"article","venue":"","topic":"Ferroelectric and Negative Capacitance Devices","field":"Engineering","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Field-programmable gate array; Computer science; Inference; Parallel computing; Computer architecture; Embedded system; Artificial intelligence","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":[0.0005211264,0.001179589,0.0003806839,0.0005238069,0.0003755025,0.0009270762,0.001777804,0.0004488723,0.01221924],"category_scores_gemma":[0.002005877,0.0005388685,0.0008475528,0.0003860365,0.000477026,0.001304197,0.0006767603,0.001040416,0.002694041],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001562378,"about_ca_system_score_gemma":0.001336265,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007611597,"about_ca_topic_score_gemma":0.01553168,"domain_scores_codex":[0.9996524,0.00005081256,0.00002125036,0.00008485402,0.0001140829,0.0000767178],"domain_scores_gemma":[0.9993598,0.00029047,0.00003731828,0.0001267573,0.0001551743,0.0000305208],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0009527041,0.0003161953,0.007800072,0.0008120005,0.0002214702,0.0007489487,0.0003896577,0.5533423,0.06619643,0.03170949,0.04340359,0.2941072],"study_design_scores_gemma":[0.00007829932,0.00006877939,0.0005618449,0.00003844922,0.00002422518,0.00006420023,0.00004337655,0.9436744,0.03749922,0.00724377,0.01068258,0.00002081184],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.149816,0.0005474623,0.7126578,0.0004364379,0.0003445187,0.000182297,0.00165687,0.1079469,0.02641174],"genre_scores_gemma":[0.5757475,0.0002307787,0.4014535,0.0003719716,0.00004135894,0.0001773968,0.003312026,0.00948099,0.00918455],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01221924,"threshold_uncertainty_score":0.04087734,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03374337786329912,"score_gpt":0.2445978066192321,"score_spread":0.210854428755933,"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."}}