{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001556151,0.0001172831,0.0001468838,0.0001070472,0.000232852,0.00003412917,0.0005566154,0.00005455955,0.000003812448],"category_scores_gemma":[0.00000570386,0.0001072068,0.00008610828,0.0008545896,0.00003549621,0.000324072,0.00005859112,0.0001147506,0.00002335328],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003013544,"about_ca_system_score_gemma":0.00002155706,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004122253,"about_ca_topic_score_gemma":0.00001928285,"domain_scores_codex":[0.9990567,0.00003464077,0.0002508069,0.0002726143,0.0001869121,0.0001983189],"domain_scores_gemma":[0.9990764,0.0002145808,0.0001102271,0.0004236279,0.0001236686,0.00005145696],"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":[0.00004461972,0.00008965079,0.00006923311,0.00008069134,0.0000271932,6.020183e-7,0.001924764,0.1141251,0.001546406,0.3201458,0.01132916,0.5506168],"study_design_scores_gemma":[0.00125832,0.0001711136,0.001767706,0.00006109999,0.00001503189,0.000003964793,0.0009333775,0.9232404,0.008874895,0.03238139,0.03095213,0.0003405371],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01035256,0.0000665816,0.9769306,0.0106273,0.0002424825,0.0006378083,0.00001768291,0.0004255479,0.0006994037],"genre_scores_gemma":[0.7803882,0.0002221296,0.2177573,0.0006897247,0.0003108231,0.000190076,0.0002152649,0.00002335085,0.0002031133],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8091154,"threshold_uncertainty_score":0.4371765,"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."}}