[Seok-Ho Seo (M.S. candidate), Dong-Hoon Kim (M.S.), Beom-Jin Kim (M.S.), Seung-Woo Park (M.S. candidate), Professor Shinhyun Choi]
KAIST EE professor Shinhyun Choi and his research team announced on 16th that they successfully developed a synthetic synapse transistor, whose durability allows it to perform repetitive operations with high reliability. The developed transistor, which emulates the behaviors of a neurotransmitter in a human brain, can also increase the learning accuracy of artificial intelligence (AI) without requiring additional circuits. By not only exploiting the structure utilized in conventional NAND Flash memory but also successfully coping with its low durability, Professor Choi and his research team were able to develop the transistor that can reliably play the role of a synapse.
This research accomplishment was published in the October editorial of Nature communications under the title of “The gate injection-based field-effect synapse transistor with linear conductance update for online training”, in which Seok-Ho Seo (M.S. candidate), Dong-Hoon Kim (M.S.), Beom-Jin Kim (M.S.), and Seung-Woo Park (M.S. candidate) were all listed as co-first authors.
KAIST EE M.S. candidate Seok-Ho Seo stated that “In addition to this research project, I want to keep on developing novel device technologies for neuromorphic computing.”
This research was supported in part by National Research Foundation of Korea, Korea Evaluation Institute of Industrial Technology, National NanoFab center, and Samsung Science & Technology Foundation.