{"id":"W4385570702","doi":"10.18653/v1/2023.findings-acl.580","title":"AutoMoE: Heterogeneous Mixture-of-Experts with Adaptive Computation for Efficient Neural Machine Translation","year":2023,"lang":"en","type":"article","venue":"","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs","keywords":"Machine translation; Computer science; Computation; Translation (biology); Natural language processing; Artificial intelligence; Linguistics; Artificial neural network; Speech recognition; Algorithm; Philosophy; Chemistry; Messenger RNA","routes":{"ca_aff":true,"ca_fund":true,"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":[0.001783543,0.001494321,0.001901058,0.0010108,0.0009611717,0.00138273,0.002824929,0.002292083,0.009032507],"category_scores_gemma":[0.004191389,0.000951261,0.001311742,0.001602582,0.0006192143,0.002781089,0.003690964,0.002418035,0.00485694],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007192375,"about_ca_system_score_gemma":0.001303285,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007516551,"about_ca_topic_score_gemma":0.01559796,"domain_scores_codex":[0.9990935,0.0003738796,0.00004265323,0.0002139988,0.0001534844,0.0001224794],"domain_scores_gemma":[0.9990331,0.0005362651,0.00003595314,0.0001795944,0.0001563109,0.00005880701],"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":[0.000970303,0.0002729552,0.0006956041,0.0001966982,0.0003671972,0.0003179835,0.0002158606,0.233266,0.007622818,0.01397517,0.0256038,0.7164955],"study_design_scores_gemma":[0.00004656747,0.00004448099,0.00008462489,0.000008275832,0.00002148535,0.00003557746,0.0000229356,0.9874557,0.001855777,0.008129827,0.002282976,0.00001183848],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.008140333,0.0007294731,0.9817882,0.0002404967,0.0001554629,0.00007877551,0.0001614735,0.006471443,0.002234358],"genre_scores_gemma":[0.2312632,0.000560111,0.7506085,0.0005918404,0.0003021979,0.0005085828,0.00169451,0.001239431,0.0132317],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009032507,"threshold_uncertainty_score":0.03021675,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02202406173855962,"score_gpt":0.2846880683209103,"score_spread":0.2626640065823506,"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."}}