In January 2021, Noh Yung-kyun, a professor in the Department of Computer Science at Hanyang University (HYU), developed a “ liver disease classification technology” using machine learning with Ahn Chul-ha, a professor of gastroenterology at the Mayo Clinic in the United States. Mayo Clinic is the best general hospital selected by the U.S. News & World Report as the “best hospital in the U.S.” for seven consecutive years. The joint study by professor Roh and Mayo Clinic was published in the American society “Ail River Meeting” in November of the same year and was selected as a “poster of differentiation.”
Q. Can you explain the scientific principle of the ʻLiver Disease Classification Technologyʼ you invented?
A. Mayo Clinic asked me to distinguish between alcoholic and non-alcoholic liver disease through machine learning. In particular, alcoholic hepatitis and nonalcoholic liver disease, cholangitis, which is a disease that causes inflammation in the bile duct that sends bile from the liver to the duodenum, are different diseases, but it is difficult to tell the difference because it has clinically similar symptoms, such as upper abdominal pain, abnormal bilestatic liver levels, and inflammatory reactions throughout the body. In addition, liver disease caused by obesity is also not easy to distinguish from alcoholic liver disease. To categorize these diseases, I learned and analyzed the vast amount of data from Mayo Clinic through machine learning.
Q. As an intellectual at the forefront of the Artificial Intelligence field, do you have any visions or goals?
A. The first thing I want to do is wrap up the book that I am currently writing so that it can be published as a good book. I have already been writing a book with my co-author professor for three to four years, and it has been said that we should write a complete book rather than publishing it quickly, and I hope I will have time to explain the content after the publication of the book. The second is to produce excellent graduates in our lab. I have a number of implications here for being a good student. Academic ability should be excellent, but it is not common for students to have such ability in advance. The laws of physics working in machine learning and the laws of Physics dealt with in real life are actually different areas, and because of these factors, many people felt high entry barriers to Machine Learning in the early days. So the great student I want to talk about should be a student with the ability to absorb the deep part of Machine Learning.
Q. Please give your advice to the future of machine learning and HYU students who are interested in it.
A. This is my personal opinion, but artificial intelligence will not develop as much as it has now. There are trends in every field, which do not last forever, and exaggerated bubbles tend to burst. However, one thing that is clear is that processing data within this field itself will have to be studied in the future.
Basically, machine learning will also become necessary knowledge in the future because the accumulated knowledge system is solid, and other academic fields using it will continue to emerge.
Learning, as the fundamental principle system in the field of machine learning, may seem boring to some extent and may be slower than people studying other applied disciplines in speed, but I think it is a wiser way for their bigger future. You have to keep thinking, “If I were the first person to invent this principle, how would I have come up with this?”