SK Hynix Declares Development of Semiconductors that Think Like Brain

SK Hynix will develop a future semiconductor device that resembles the structure of a human brain in partner with Stanford University of the US. 

SK Hynix announced on October 13 that SK Hynix signed an agreement on the joint research of artificial neural network semiconductor devices based on ferroelectric materials at Stanford University in California of the US in the afternoon of October 12 (local time). Lam Research and Versum Materials will join the collaboration project. The former is a semiconductor equipment manufacturer while the latter produces semiconductor materials.  

SK Hynix expects this program to become a turning point in the development of a neuromorphic chip. A neuromorphic chip is a computing system loaded with an artificial neural network semiconductor device that can think and judge like a human brain. 

Current computer systems receive a datum and save it in a memory semiconductor after a CPU and reanalyze it. But a neuromorphic chip enables a computer system to intuitively analyze and save multiple data at the same time like human beings.

SK Hynix expects that the development of an artificial neural network and a neuromorphic chip will ultimately lead to a new-concept computing system with both memory semiconductor functions and system memory semiconductors’ calculation capabilities. 

“This joint research is expected to accelerate the development of an artificial neural network semiconductor device by making the most of each participant’s strengths such as devices, processes, equipment, materials and design,” said Hong Sung-joo, vice president of SK Hynix and director of the company’s Future Technology R&D Center.