The South Korean government will provide 200 graphics processing units next year to university research institutes taking on difficult scientific and technological problems in fields including semiconductors, biotechnology and space. The plan is to help these institutes experiment with and develop innovative technologies using artificial intelligence and train researchers over the next six years.
The Ministry of Science and ICT is moving to provide a combined 200 GPUs to 15 university research institutes starting next year through its AI-based university science and technology innovation program, according to the science and technology community on the 25th. The government launched the program this year to build a sustainable research ecosystem for university science and AI, and began soliciting applications from institutes and research projects in April.
The government plans to use the program to help researchers tackle unsolved scientific problems and to train specialists in both AI and science and technology. While the government's K-Moonshot Project addresses 12 science and technology missions deemed nationally important, this program takes a bottom-up approach in which researchers themselves identify the problems they want to solve. The research period runs six years, with each project expected to receive 2 billion won ($1.4 million) a year, or about 12 billion won ($8.4 million) in total.
The university institutes will pursue missions such as developing treatments for liver disease and using AI to predict mental illness and suicide risk. One approach is to use AI to analyze large-scale data including brain waves, hormones and clinical records to identify the causes and progression of depression and bipolar disorder, and to predict suicide risk associated with mental illness. In semiconductors, domestic researchers plan to develop original technology for three-dimensional integrated circuits that reduces heat generation by using AI to optimize materials, devices, packaging and architecture. The researchers aim to secure core competitiveness in next-generation 3D IC technology. Other efforts include using AI to determine the particle identity of dark matter, whose nature has remained unknown for about 90 years since its existence was first proposed in the 1930s, and using AI to design the molecular environments in which chemical reactions take place.
The institutes will work on their projects with the Gwangju Institute of Science and Technology (GIST) serving as the central hub. GIST was selected in July as the lead organization for building the program's central infrastructure and has begun constructing and operating AI research facilities for shared use by universities. GIST will build a shared high-performance computing and AI infrastructure (HPC-AI) to meet science and technology research demand and will oversee the nationwide shared-use service.