{"id":"W4401211627","doi":"10.1109/isca59077.2024.00078","title":"Pre-gated MoE: An Algorithm-System Co-Design for Fast and Scalable Mixture-of-Expert Inference","year":2024,"lang":"en","type":"article","venue":"","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":36,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"","keywords":"Computer science; Inference; Scalability; Algorithm; Parallel computing; Artificial intelligence; Database","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":[0.001081032,0.0009299993,0.001014849,0.0005127347,0.0004150437,0.001063,0.002883253,0.001216879,0.007197674],"category_scores_gemma":[0.004676475,0.000735868,0.0008087476,0.0004319742,0.0006816935,0.001966523,0.002253955,0.002189545,0.00205454],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006603886,"about_ca_system_score_gemma":0.001700867,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003542888,"about_ca_topic_score_gemma":0.008512551,"domain_scores_codex":[0.9993973,0.0001401298,0.00003809742,0.0001525306,0.0001924227,0.00007953418],"domain_scores_gemma":[0.9989946,0.0004756713,0.00004662,0.0001843083,0.0002237788,0.00007505922],"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":[0.0005986574,0.0002485384,0.002033401,0.0003243702,0.0002355986,0.0002839697,0.0002401718,0.4454682,0.02624782,0.02395102,0.008364744,0.4920035],"study_design_scores_gemma":[0.00001978681,0.00004813915,0.00008660553,0.000006436862,0.00001030919,0.00004334794,0.000009613122,0.9893041,0.004361947,0.004099499,0.002001125,0.000009165481],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.003758382,0.0001135025,0.9918583,0.00006313792,0.0000291551,0.00006921354,0.00003337587,0.00340164,0.0006732521],"genre_scores_gemma":[0.2281601,0.0002014999,0.7662956,0.0003862449,0.00003991706,0.0003939812,0.0004241457,0.0008103398,0.003288175],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.007197674,"threshold_uncertainty_score":0.02407861,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03050168526136428,"score_gpt":0.3411675755153544,"score_spread":0.3106658902539901,"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."}}