[Asia Economy Reporter Junho Hwang] Domestic researchers have developed an artificial intelligence that can accelerate the speed of physical property analysis for developing next-generation semiconductor materials. By estimating the physical properties from big data of spin structure images of magnetic materials, the research speed for next-generation semiconductor development is expected to increase.
The research team led by Dr. Heeyoung Kwon and Dr. Junwoo Choi of the Spin Fusion Research Group at the Korea Institute of Science and Technology (KIST), in collaboration with Professor Changyeon Won's team at Kyung Hee University, announced on the 17th that they have developed such an AI, and the related research paper was recently published in the international journal Science Advances.
The team applied machine learning algorithms to the AI to enable it to predict the physical properties of magnetic materials. First, they trained it with existing magnetic domain images. Then, when presented with new magnetic domain images, the AI estimated the magnetic properties of the material. As a result, by inputting electron microscope images of magnetic materials, it became possible to estimate the magnetic properties of the material in real time. When comparing the AI's estimations with actual observed data, the error was within about 1%, indicating very high estimation accuracy.
Dr. Heeyoung Kwon of KIST stated, "We have presented a new path for how AI technologies can be utilized to analyze the characteristics of magnetic domains," adding, "Using such AI technology to analyze magnetic systems offers a new research method that strengthens the connection between experiments and theory, and it is expected to lead to the expansion of a new research field that fuses AI technology with pure science research."
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
