Korea's academic level for 'computer vision artificial intelligence (AI)' technology was ranked third~fourth following the U.S. and China.
Computer vision AI technology performs face and object recognition. In addition to the automobile-related autonomous driving technology, which is a major export item, the application for AI fields are diverse, such as robots, medicine, security, and games.
However, industry continuously expressing that there is a shortage of talents. The situation requires a government policy to solve the problem.
According to the industry on the 28th, Korea ranked third after analyzing the result of countries that submitted and selected papers at the 'Conference on Computer Vision and Pattern Recognition (CVPR) 2021’, which held last month.
CVPR 2021 is one of three major international academic conferences in the computer vision field. This year’s conference had a total of 12 sessions, and selected 1601 papers.
The selected papers were submitted not only by academia such as Seoul National University, Korea Advanced Institute of Science and Technology (KAIST), Yonsei University, Korea University, and Sogang University, but also by companies such as Naver, Kakao, and LG.
The country with the most selected papers was China, accounting for 42%. The U.S. ranked second with 23.7%.However, the U. S. ranked as number one in the world, considering the quality differences with Chinese papers.
Seoul National University professor Kyoung-mu Lee, who served as a member of the CVPR 2021 Awards Committee, said, “Korea has been ranked 3rd and 4th not only in CVPR, but also in the International Conference on Computer Vision (ICCV) and the European Conference on Computer Vision (ECCV) in recent years. Although there is a gap with the 1st and 2nd places, Korea is growing rapidly considering proportion to our population.”
However, the industry agrees that there is a shortage of talents in the field of computer vision AI. PhD-level talents are needed in the field. However, the fact is that there is lack of AI talents in Korea since nurturing program only started recently in 2016, when the Se-dol Lee vs. AlphaGo go match was held.
Firstly, there is a discrepancy in expected qualification between academia and companies, which raised an opinion that number of quotas should be increased and curriculum should be diversified simultaneously. Secondly, it is necessary to expand AI-related facilities to cultivate practical skills comparable to work facilities in companies.
A manager at a large AI company was concerned and said, “A few global companies openly distribute their data sets, but the reality is that universities cannot use the data sets due to its insufficiently equipped facilities. There are domestic industrial academic projects and internships are alternative option; Due to the lack of well-equipped domestic facilities, our excellent talents are flocking to overseas internships.”
The government needs to consider supporting development and usage of cloud systems in order to nurture excellent AI talents. Considering the pace of AI development, it is not a problem that universities can solve by just providing better facilities.
Professor Lee said, “Most universities have reached the limit of facility expansion. There are universities that lack fund for investment, or universities that lack space even when they have the fund. It came to a point where even electricity consumption cannot be afforded. The cloud system is an alternative, but the high cost stands as an issue.
Expansion of industry-academic projects between domestic companies and universities is also necessary. Universities can nurture AI talents, and domestic companies can preempt these excellent talents, which is a win-win situation. It is also necessary to support and encourage the concurrent positions of professors in AI-related companies. There are advantages that allow universities to hire well-qualified professors, and students to gain practical experiences early with professors at their companies.
Professor Lee said, “Although the government is continuously funding budgets to nurture AI talents and producing positive academic results, a detailed support is necessary for universities to produce qualified manpower required by the industry.”
By Staff Reporter Jin-hyung Park