{"id":"W4414260593","doi":"10.1145/3767334","title":"SuperBench: A Proactive Validation System for Improving Reliability of Cloud AI Infrastructure","year":2025,"lang":"en","type":"article","venue":"ACM Transactions on Computer Systems","topic":"Software System Performance and Reliability","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"ca_institutions":"Microsoft (Canada)","funders":"","keywords":"Testbed; Cloud computing; Benchmark (surveying); Validator; Reliability (semiconductor); Fault tolerance; Root cause","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":[0.0008776153,0.0002855792,0.0005310242,0.0003015475,0.0003282016,0.0001755502,0.001201177,0.0002463108,0.000001345184],"category_scores_gemma":[0.00004622605,0.0002383263,0.0002583007,0.0007664955,0.00006443061,0.0006753685,0.00004872258,0.0002688937,0.000005467545],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003784215,"about_ca_system_score_gemma":0.0003025142,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002074411,"about_ca_topic_score_gemma":0.000005250609,"domain_scores_codex":[0.9974731,0.0002404127,0.0008534361,0.000755712,0.0003510893,0.0003262358],"domain_scores_gemma":[0.9965851,0.0005611692,0.0002379621,0.001874325,0.0006652563,0.00007621131],"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":[0.0006872669,0.001592738,0.01209909,0.0325093,0.001179017,0.000007446297,0.007404676,0.6236452,0.002762303,0.03618328,0.006535175,0.2753945],"study_design_scores_gemma":[0.002515967,0.0009330572,0.00471758,0.001200581,0.0001594797,0.00005206751,0.0005150369,0.9573158,0.02663489,0.001481628,0.003753668,0.0007202569],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0255962,0.00004940762,0.9611426,0.0003040765,0.01049824,0.001885525,0.00005080758,0.0004091377,0.00006403676],"genre_scores_gemma":[0.9519133,0.000001789542,0.04733555,0.00005142594,0.0002436727,0.0003694415,0.000008049802,0.00001338932,0.00006338938],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9263171,"threshold_uncertainty_score":0.9718666,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.008737100156824773,"score_gpt":0.2430366704539447,"score_spread":0.2342995702971199,"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."}}