RaPiD: AI Accelerator for Ultra-low Precision Training and Inference.
Swagath VenkataramaniVijayalakshmi SrinivasanWei WangSanchari SenJintao ZhangAnkur AgrawalMonodeep KarShubham JainAlberto MannariHoang TranYulong LiEri OgawaKazuaki IshizakiHiroshi InoueMarcel SchaalMauricio J. SerranoJungwook ChoiXiao SunNaigang WangChia-Yu ChenAllison AllainJames BonannoNianzheng CaoRobert CasatutaMatthew CohenBruce M. FleischerMichael GuillornHoward HaynieJinwook JungMingu KangKyu-Hyoun KimSiyu KoswattaSae Kyu LeeMartin LutzSilvia M. MuellerJinwook OhAshish RanjanZhibin RenScot RiderKerstin SchelmMichael ScheuermannJoel SilbermanJie YangVidhi ZalaniXin ZhangChing ZhouMatthew M. ZieglerVinay ShahMoriyoshi OharaPong-Fei LuBrian W. CurranSunil ShuklaLeland ChangKailash GopalakrishnanPublished in: ISCA (2021)
Keyphrases
- structured prediction
- artificial intelligence
- machine learning
- high speed
- neural network
- ai systems
- inference process
- training phase
- training algorithm
- bayesian inference
- intelligent systems
- case based reasoning
- knowledge representation
- expert systems
- test set
- training process
- online learning
- real time
- ai technologies