{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006177212,0.0001666433,0.0002292726,0.0001275098,0.0001159552,0.0002771913,0.0005313503,0.0001072283,0.00001746255],"category_scores_gemma":[0.00007074505,0.0001340244,0.00003525388,0.0003396123,0.00004242862,0.0004751949,0.0001088754,0.000174618,0.00001009157],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004638851,"about_ca_system_score_gemma":0.000177561,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007403388,"about_ca_topic_score_gemma":0.00001291804,"domain_scores_codex":[0.9984246,0.0001935381,0.0003084158,0.0005412326,0.000240165,0.0002920317],"domain_scores_gemma":[0.9986439,0.0005175147,0.000060363,0.0004663685,0.0001550186,0.0001567824],"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":[0.00002589362,0.00008224042,0.001922128,0.001836219,0.00005980507,0.00002318368,0.00787784,0.002370478,0.001477304,0.1056306,0.004633471,0.8740609],"study_design_scores_gemma":[0.0001420253,0.000393298,0.0004186154,0.0001617325,0.000003390343,0.00002035572,0.00007268619,0.9929066,0.002640847,0.0004773291,0.002604591,0.0001584854],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.001645753,0.0005655327,0.9950494,0.0005741717,0.0003263905,0.0005792165,0.00001154823,0.0006476701,0.0006003815],"genre_scores_gemma":[0.4507098,0.00001483431,0.5481755,0.0001217776,0.00008570724,0.00006446984,0.000009788971,0.00001952094,0.000798562],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9905362,"threshold_uncertainty_score":0.5465358,"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."}}