A contest in the field of artificial intelligence was one of the crowd-pullers at the National Exhibition of Achievements in Information Technology 2019 last November.
The artificial intelligence institute under the faculty of information science of Kim Il Sung University outdid all others in the programming contest, which was held in such sectors as machine translation, voice, letter, face and fingerprint recognition, and car license number identification.
“The secret of our success is mainly the application of the deep leavning, which is one of the hottest technologies worldwide,” said Choe Chol, director of the institute.
According to him, a fierce competition is being staged in the AI field to develop recognition systems almost on a par with human intelligence by using the deep leavning. However, the application of the technology is unthinkable apart from the establishment of a huge database.
The institute has been engaged in the development and upgrading of Korean letter and voice recognition technologies for over 20 years and now takes the dominant position in this field on the domestic front.
Its Korean letter recognition program can read printed and cursive letters with the accuracy of over 99.5 and 95 percent respectively, while the Korean voice recognition program has a recognition rate of over 98 percent.
“As we have laid such a firm foundation, we could win the car number identifying contest held for the first time in the country,” said section chief Kim Sun Dol.
This program is said to be able to correctly identify the license plate numbers of cars running at high speed.
The institute’s English-Korean and German-Korean translation programs also show the high level of its machine translation technology.
Last year, its software development process was awarded the nation’s top-level quality management system certification, and on that basis the institute developed an integrated retrieval system and an integrated electronic procedure system.
“It is a vital matter in information retrieval to set up a big data retrieval system as soon as possible,” said Pak Tae Hyok, a section chief of the institute.
The integrated retrieval system can collect, store and manage public information offered through websites before analysing and retrieving the data.
The integrated retrieval system and electronic procedure system were appreciated as best IT products last year.
The institute was also successful in the development of network security management systems, web pages and mobile apps.
It was chosen as the best IT business in 2016 and 2017 and one of the top 10 IT businesses in 2018 and 2019.
“Our research effort has focused on artificial intelligence this year and we’ll go on with this effort and bring out more AI-related products, while developing basic technologies to gain a competitive edge in the world,” said Choe Chol.