Specialized hardware accelerators designed to significantly speed up the calculations required for ML tasks.
Generality: 821
