July 30, 2022
Improving AI Using Semiconductor Materials
It is well known that vacuum deposition is used in the semiconductor industry to build chips layer by layer. Many may also know that semiconductors make use of ferroelectric materials. However, what many people don't know is that now these same materials and processes will be used to improve AI computation through hardware architecture. Applications of AI are becoming more widespread and with this comes the need for more computational power and better energy efficiency to sustain that extra power. Not to mention that as AI becomes more popular it also becomes slower because it takes longer for data to travel to the user from the distant data centers where much of the computation takes place. Consequently, in an effort combat massive computational and energy costs, engineers from the University of Tokyo are applying the use of hafnium-oxide ferroelectric materials (similar to those used to build semiconductors) on physical reservoir computing (PRC). Read more here.