Exploration and design of super functional materials and devices using evolutionary learning and artificial neural networks
일 시 : 2021년 11월 11일 목요일 16:00
연 사 : 이 인 호 박사 (표준과학연구원)
장 소 : 온라인 진행 ( https://us02web.zoom.us/j/81101834598?pwd=RkZIdXJLL1lOWXBjamc2Ylg3VUtRZz09 )
HOST : 조 준 형 교수님
초 록
Using both artificial neural networks and evolutionary learning methods, we have explored various super functional semiconductors, superconductors, and topological superconductors. In addition, for devices, we have designed the frequency-selective meta-surfaces that can improve the quality of 5G communication. For each study, we first define a figure-of-merit and then search the corresponding conformation space. We thoroughly utilized the three key elements of representation, optimization, and evaluation as found in common machine learning. Although not all designed cases have been experimentally implemented, we present that in some cases they can be implemented experimentally.