Domestic researchers are contributing to the expansion of the Korean AI ecosystem through research on the Korean AI research and development project 'Exobrain'.

The Electronics and Telecommunications Research Institute (ETRI) announced on the 30th that the Korean AI 'Exobrain' project has concluded its 10-year journey. This project achieved the following results: △490 domestic and foreign papers, △362 domestic and foreign patent applications, △8.5 billion won in technology transfer, 13.3 billion won in commercialization, and △2 companies listed on KOSDAQ.

Exobrain is a project that was conducted over a period of 10 years in three stages from May 2013 to February of this year, hosted by ETRI, Saltlux, and KAIST.

The Exobrain research team has been developing Korean AI technologies, including △deep question-answering technology that finds answers to natural language questions from searched documents, △Korean language analysis technology, △deep learning language model-based application technology, and △knowledge base-based inference technology.

As an achievement, in November 2016, he won the scholarship quiz <Showdown! Exobrain> against human quiz kings. In addition, AI technology and machine learning data were distributed for the first time in domestic R&D projects through open API and data service portals starting in 2017.

The research team has so far released about 18 types of language processing technologies, 14 types of learning data, and 2 types of language models. As of the end of 2022, the released data has been used in more than 61 million cases in 2,349 institutions.

In 2019, the Korean language model KorBERT was developed and released for the first time in Korea, with performance evaluated to be 4.5% better than Google's language model BERT.

Among the institutions that participated in the study, Saltlux was listed on KOSDAQ in July 2020, and Minds Lab, a joint institution of the first division, was listed in November 2021.

The ETRI research team announced that since April of last year, they have been conducting research on the development of artificial intelligence SW that can support decision-making by experts in various fields by expanding the explainability of natural language processing technology as a follow-up project to the Exobrain project.

ETRI said, “Unlike ChatGPT, which has good generality, we expect to be able to show differentiated results by providing natural language-based explanations (correct answers) specialized for specific domains such as law and finance, along with reliable grounds for those explanations.”

IITP PM Lee Hyeon-gyu said, “Exobrain is a representative technology that is the result of a national project and is being used the most in the commercialization market, and has made a significant contribution to the development of domestic artificial intelligence technology. “It seems desirable to pursue it as a mission-oriented project that will enhance current performance or insufficient functions in the future,” he said.

ETRI’s Dr. Baek-Man Bae, who is in charge of the Exobrain project, said, “We will work to ensure that the results of the Exobrain project, accumulated over the past 10 years, are not wasted but are continuously utilized, and we will work hard to develop advanced technologies in Exobrain technology through follow-up projects with the project members.”