[Reporter Hae Chan Kim of Veritas Alpha] Professor Sung-rho Yoon’s research team (Seoul National University (SNU) Electrical & Computer Engineering (ECE) Department and Interdisciplinary Program in Artificial Intelligence (IPAI)) reported that they developed an AI-based technology that can predict molecular properties more than 1,000 times faster than previous methods. It is predicted that this technology can contribute to improving the efficiency of molecular property predictions during development of advanced materials, boosting the innovation cycle of advanced material developments.

The research team explained the operation of the newly developed artificial intelligence-based ultra-fast molecular property prediction technology by drawing an analogy to an artificial intelligence recognizing images. They stated, "If there is an artificial intelligence trained only on high-quality pictures painted by professional artists, it will recognize high-quality pictures well. However, when it sees a child's sketch of a cat drawn with a pencil for the first time, it may find it difficult to recognize that it is a cat. The newly developed artificial intelligence has been trained to recognize common features between realistic pictures and those drawn with a pencil, enabling it to effectively recognize various types of drawings." Such advancements allow for precise property predictions of new molecules, even when using molecular information obtained from simple and quick methods, eliminating the time and cost-intensive methods traditionally used for predicting the characteristics of new molecules.

Professor Sung-rho Yoon stated, “In recent years, research on advanced materials using artificial intelligence has been spreading. AlphaFold of Google DeepMind and GNoME(artificial neural network for material searching) are prominent examples. With the utilization of generative artificial intelligence, there is an emerging trend of complex pipelines that generate candidate materials and experimentally validate them. As a result, it is anticipated that artificial intelligence will play an increasingly significant role in the revolutionary development of novel materials that humanity has dreamed of. I appreciate the support from the National Research Foundation of Korea (NRF) and the Institute of Information & Communications Technology Planning & Evaluation (IITP), who spared no effort in facilitating pioneering interdisciplinary research."

Meanwhile, this research was conducted with the support of the Basic Research Program, BK21 Project in the field of science and technology by the National Research Foundation of Korea, and the Artificial Intelligence Graduate School Program by the Korea Information Society Development Institute (IITP). The findings were published in the international academic journal, Nature Computational Science, on December 5, 2023.

Professor Sung-rho Yoon’s research team of SNU ECE department and IPAI reported that they developed AI-based technology that can predict molecular properties 1,000 times faster than previously used methods (picture from SNU)

Source: https://www.veritas-a.com/news/articleView.html?idxno=484890

Translated by: Jiyong Yoo, English Editor of the Department of Electrical and Computer Engineering, cyoo7@snu.ac.kr