{"id":"W4385571590","doi":"10.18653/v1/2023.findings-acl.180","title":"Operator Selection and Ordering in a Pipeline Approach to Efficiency Optimizations for Transformers","year":2023,"lang":"en","type":"article","venue":"","topic":"Machine Learning and Algorithms","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Canada First Research Excellence Fund","keywords":"Computer science; Commutative property; Pipeline (software); Operator (biology); Quantization (signal processing); Inference; Transformer; Theoretical computer science; Artificial intelligence; Algorithm; Mathematics; Programming language; Engineering","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000227959,0.00005652818,0.00006669582,0.0001862967,0.0001059854,0.00007616235,0.0001131697,0.00001942565,0.000001927145],"category_scores_gemma":[0.00003881315,0.00004923596,0.00001299,0.001167695,0.00000421366,0.0001161002,0.0000277549,0.00005093291,0.000004134703],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000009869848,"about_ca_system_score_gemma":0.00002290465,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005105308,"about_ca_topic_score_gemma":0.00002029438,"domain_scores_codex":[0.9994426,0.0000130566,0.0000980103,0.0002190763,0.00006250505,0.0001646888],"domain_scores_gemma":[0.999836,0.00002983266,0.000007260279,0.00005352573,0.00002519337,0.00004823174],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000003256321,0.00005631795,0.0002894498,0.00002538211,0.000002373402,3.001222e-7,0.001889701,0.9120137,0.0003461014,0.007397801,0.0003792763,0.07759636],"study_design_scores_gemma":[0.0002286586,0.00004493007,0.0002378468,0.000003814577,7.537462e-7,0.000002404377,0.00006760855,0.9982131,0.000127112,0.0000307949,0.0009711889,0.00007180243],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.009800292,0.000005215741,0.9870954,0.001048361,0.00004923229,0.0002410524,6.467948e-7,0.0002061693,0.001553603],"genre_scores_gemma":[0.4667571,0.000008780235,0.531399,0.0001853796,0.00003434409,0.0001072541,0.000005573212,0.000009050771,0.001493449],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.4569569,"threshold_uncertainty_score":0.2007784,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01306476756127116,"score_gpt":0.263144824771561,"score_spread":0.2500800572102899,"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."}}