{"id":"W4403364019","doi":"10.48550/arxiv.2410.07381","title":"Tally: Non-Intrusive Performance Isolation for Concurrent Deep Learning Workloads","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Semiconductor materials and devices","field":"Engineering","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; VMware","keywords":"Computer science; Deep learning; Isolation (microbiology); Artificial intelligence; Computer architecture","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0001059161,0.0003025236,0.0003010804,0.0001621128,0.00008339238,0.0001156767,0.0002613647,0.000268727,0.0001171156],"category_scores_gemma":[0.00001118415,0.0003505475,0.0001488611,0.00016256,0.00003428728,0.0001333046,0.0003246289,0.000485754,0.0001545137],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001949724,"about_ca_system_score_gemma":0.00003819705,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001540938,"about_ca_topic_score_gemma":0.00001103307,"domain_scores_codex":[0.9989925,0.00001792929,0.0001886398,0.000476663,0.00004071382,0.0002835657],"domain_scores_gemma":[0.9994747,0.0000556789,0.00008580281,0.000234394,0.00008325042,0.00006615451],"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":[0.00002432949,0.00001025588,0.0006037495,0.001596143,0.0001755104,0.00001274424,0.0004190814,0.987015,0.007492373,0.0009926726,0.0002459784,0.00141213],"study_design_scores_gemma":[0.0003222383,0.0000441944,0.0004593901,0.0004705284,0.0001755368,0.000001638538,0.0001588514,0.9910367,0.00358186,0.001135188,0.00213655,0.0004773565],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9769046,0.0008417941,0.01596904,0.000005844508,0.003986245,0.0003820666,0.00002558725,0.0004471646,0.001437715],"genre_scores_gemma":[0.9981683,0.0006522753,0.00008736822,0.000008505564,0.0002829707,0.00000577005,0.00009587105,0.00006006609,0.0006388983],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02126373,"threshold_uncertainty_score":0.9998947,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0357364504144257,"score_gpt":0.1773598276114469,"score_spread":0.1416233771970212,"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."}}