{"id":"W4387951242","doi":"10.1109/iiswc59245.2023.00026","title":"Tale of Two Cs: Computation vs. Communication Scaling for Future Transformers on Future Hardware","year":2023,"lang":"en","type":"article","venue":"","topic":"Advanced Neural Network Applications","field":"Computer Science","cited_by":18,"is_retracted":false,"has_abstract":true,"ca_institutions":"Advanced Micro Devices (Canada)","funders":"","keywords":"Computer science; Profiling (computer programming); Computation; Transformer; Scaling; Distributed computing; Computer engineering; Algorithm","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":[0.00239322,0.001031617,0.0006601476,0.0007216635,0.0007163662,0.002258062,0.001896961,0.001086126,0.006249504],"category_scores_gemma":[0.01520471,0.0006605202,0.0008557102,0.0009216422,0.001746747,0.008088369,0.002331955,0.00360851,0.001092574],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001611413,"about_ca_system_score_gemma":0.001984543,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004946441,"about_ca_topic_score_gemma":0.007617089,"domain_scores_codex":[0.9985714,0.0003272881,0.00006008654,0.0002896566,0.0005145395,0.0002370445],"domain_scores_gemma":[0.9933142,0.002695401,0.0002565946,0.002404156,0.0007789376,0.0005507515],"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":[0.001282859,0.0004874263,0.01769001,0.0004265035,0.000208699,0.0004048876,0.0007775332,0.5897259,0.03053465,0.09213094,0.02793713,0.2383934],"study_design_scores_gemma":[0.00006464434,0.0003376498,0.002891267,0.00004606527,0.00006036121,0.0001619013,0.0003849767,0.9280546,0.01431534,0.04518155,0.008458223,0.00004333448],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5947563,0.002279646,0.3554921,0.0120596,0.0006388337,0.0001603676,0.0007659465,0.01010408,0.02374307],"genre_scores_gemma":[0.9121224,0.0005608877,0.08223827,0.0006112394,0.00009047202,0.00009987653,0.0005609104,0.0009328006,0.002783097],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.006249504,"threshold_uncertainty_score":0.02090675,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02227682947137752,"score_gpt":0.3116673841263964,"score_spread":0.2893905546550189,"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."}}