|Professor Park Ji-hong (n.news.naver.com)|
Park Jin-hong, a professor from the Department of Electronic, Electrical, and Computer Engineering and hisresearch team was able to develop an AI semiconductor device that can enhance the information processingcapabilities of the next-generation neuromorphic chips that mimic the learning principles of the human brain. Neuromorphic chips have garnered attention as the nextgeneration information processing chips because they process large amounts of information in parallel to minimizing power consumption.
Recently, they have conducted a research on synaptic imitativesemiconductor devices that are essential for parallel information processing and learning ability implementation of neuromorphic chips. Professor Park’s team combined P-type semiconductor channels that are suitable for strengthening linear long-term memory and N-type semiconductor channels that are suitable for weakening long-term memory and implemented mixedchannel synaptic semiconductor devices that have both linear and symmetrical long-term memory strengthening and weakening characteristics. Furthermore, a research team, which utilized the hardware artificial neural network, succeeded in improving the recognition rate of voice information patterns. They have also stated that the learning and recognition rate of voice information patterns in the hardware artificial neuralnetwork improved by more than 15% compared to the already existing artificial neural network.
<저작권자 © THE SUNGKYUN TIMES, 무단 전재 및 재배포 금지>