[Meeting with a Great Scholar, Professor Moon, Songcheon from KAIST] Question about AI in Korea2016-12-29Hit:690
[Meeting with a Great Scholar, Professor Moon, Songcheon from KAIST] Question about AI in Korea
"It is sad that the current government policy of Korea attaches too much so a single tree called Artificial Intelligence (AI) instead of the entire forest called Software (SW),"
A pioneer and an expert in computer science in Korea, Professor Moon, Song-cheon (63) from the School of Business at KAIST said this, insisting that Korea shall react differently from the "Alpha Go Shock" in an interview with the Metro Newspaper.
He said that development of any software without the infrastructure – the operating system (OS) and database (DB) engine – is just a house of cards and that Korea needs the proper national software development research institute to begin the development of software industry. This implies that the excessive attachment to only AI will just grant more opportunities for those who already dominate the operating systems and database engines. The fact that the U.K. created Alpha Go but handed it to Google of the U.S. provides some circumstantial evidence.
Professor Moon is giving lectures at the National University of Ireland and Newcastle University of the U.K. this semester. He had this interview by e-mail and over the phone. The following is a summary of his interview.
-After the match between Lee, Se-dol and Alpha Go, this wave of AI in Korea is very powerful.
"Alpha Go is originally a work by the British corporation called DeepMind. Yet, in the U.K., a press release on it was provided only three times: once by BBC, once by the Times, and once by the Guardian. It is quiet a contrast from the fact that there were thousands of news items during the match season in Korea. Moreover, the follow-up measures were also different. After the U.K. and Ireland watched this Alpha Go incident, the government realized that education on computer coding is supposed to be provided to children in the earlier stages and announced that the government will introduce a program for education on coding for elementary school children. On the other hand, in Korea, the national AI Research Institute was launched, strangely."
-Is there any problem with launching the National AI Research Institute?
"AI doesn’t even account for 10% of the entire software field. It is not logical to bet everything on AI, and this house-of-cards investment will hinder the developmental balance of the entire software field. In the software field, operating systems and the database engines are the basic infrastructural technology that make up 80% of the entire software field. The right direction is to promote the basic field instead of being attached too much to AI to resolve this impact by Alpha Go. Alpha Go from the U.K. is software created by a few members of Cambridge University. They did not receive any support by the British Government. Yet, AI in Korea has been around for over 30 years. If there was proper research, rather than Alpha Go, there would be three or four 'Super Alpha Go’s.' If the AI Research Institute does the same thing that has been done for the past 30 years, there will be meaningless studies, and it is therefore not adequate."
-How come AI is not the center of the software industry?
"There must be a soccer field for a great soccer player like Messi to play soccer. Likewise, the core software is a platform like a field. It is a field for a player to perform. The platform itself is composed of a multi-layer structure, but the most basic infrastructure is the operating system like Windows of Microsoft and the database engine like Oracle. OS and DB Engines, without these two, there is no software that will work properly, and AI like Alpha Go is not an exception."
-What happens if the AI is developed without an OS and DB engine?
"Although the U.K. created Alpha Go, since it is software that is operated under the OS and the DB engine created in the U.S., instead of proving the competence and competitiveness of the U.K., it has shown the competitiveness of the U.S. Since Alpha Go runs on the software uniquely possessed and provided by Google, it performed well in the match with Lee, Se-dol. Without the platform by Google, Alpha Go would not be able to perform so well DeepMind of the U.K., the creator of Alpha Go, has been taken over by Google of the U.S. because there was no solution for using Alpha Go in the U.K."
-Are you saying that Korea may follow what the U.K. has faced?
"Yes. Surely. Without being successful in the domestic development of OS and DB engines, all investment in the software fields will be obsolete. Without knowing that Google has its own OS and DB engine providing an autonomous ecosystem, and that Microsoft and IBM have their own OS and DB engines, you won’t be be able to see this. You should understand why they are not taking the platforms provided by other companies. Once a company takes the platform created by others, it will face the fate of being taken over by other companies just like DeepMind of the U.K. In conclusion, it is inevitable to give their achievements to others."
-Korea is a leader in IT. Won’t Korea perform better than the U.K.?
"Hardware is surely strong, but software in Korea is much weaker than in the U.K. In IT, software takes 80%, and hardware takes 20%. In the hardware field, the U.S. and Korea are dominating the global market. On the other hand, in the software market, the U.S. is dominating – it has more than 80% of the global market, and the U.K. is following the U.S. with a market share of about 10%. Korea is taking only 0.% of the market for the software industry, showing that Korea is way behind. This abnormal feature of leading in HW with no SW is the sad truth of the IT industry in Korea. The value added from the software field is so huge that Korea will never give up on this field. Korea has had ability in the HW industry for the past 40 years, and it must establish this in the SW industry as well."
-What shall Korea do?
"The error of launching the National AI Research Institute by the government needs to be rectified. The National SW Development Research Institute needs to be set up properly to begin the development of the SW industry. Some academics and some political authorities have raised their voices saying that AI represents the entire SW industry. It is now time to calm down and correct the plan carefully. There is this SW Development Research Institute in Korea, but if this research institute just brings in the imported SW to assemble or manipulate, in perspective of a platform, this institute is meaningless."
-With respect to AI, some worry about massive unemployment.
"Historically, no one doubts that machine technology has contributed to the creation of more jobs for humans instead of destroying human society. Look at the advent of the transistor – it has created countless jobs. Look at the advent of the computer. Some said that this office work will be automated to reduce jobs, but actually it was different from what was expected. AI will also follow this trend. More jobs will be created from AI – it won’t threaten conventional jobs, as we can learn from history."
-Will AI surpass humans?
"Humans have the intelligence to overcome difficulties wisely. Although AI will be continually developed, it will be quite difficult for it to surpass natural intelligence."
-You are working as the first Korean director of Euromicro.
"The limit in Asia is that there is almost no academic journal to represent to represent Asia globally. It is important to show that the level of research outcomes from Asia is equivalent to the level of research outcomes from European countries, and there needs to be a person to lead this. I think am fortunate to be able to contribute to this process."
※Professor Moon, Song-cheon is the first doctorate holder in computer science in Korea and a renowned authority in OS and DB. He graduated with a Bachelor's Degree in computer science from Soongsil University and finished his Master's Degree at KAIST. Thereafter, he taught computer science at Soongsil University and went to the U.S. to receive his doctoral degree in computer science from the University of Illinois. When he was a professor of computer science at KAIST in 1990, he developed the DB engine, IM and the distributed DB engine DIME, the most esoteric field in cutting-edge software field, after the U.S. for the second in the world to amaze academia. (I don’t understand this at all.)
He is one of only a few people involved in computers in the early 1970s like Bill Gates or Steve Jobs, and he began his career as a professor from the very young age of 24. He has continued to nurture his students for 40 years. His scope of activities covers all of Europe. He was a professor of computer science at Cambridge and Edinburgh of the U.K., and he is currently the representative of Asia at Euromicro, one of the three global IT societies. He is a professor in the College of Business at KAIST, and he gives lectures at the School of Management of the National University of Ireland and the Computer Science Department at Newcastle University of the U.K. in this fall semester.
Besides research and teaching, he contributes to society through volunteer activities. He initiated the donation marathon which involves running and receiving donations for each meter, and he has completed the marathon twice a year for 17 years with KAIST students. Moreover, he is has been involved in radio broadcasts once a week for the past 20 years for popularizing IT. He donated the entire performance fee to help the children with leukemia and victims of comfort women during the World War II.
<From Metro Newspaper>