Research Highlights

POSTECH Opens the Graduate School of Artificial Intelligence, Stepping Forward to Change the World

2020-02-06 373

As the Fourth Industrial Revolution has begun, the demand for AI skilled professionals has also increased. In the midst of this fast-changing era, POSTECH will soon open the Graduate School of Artificial Intelligence (hereinafter referred to as AI Graduate School). AI Graduate School is an initiative by the Ministry of Science and ICT of Korea in collaboration with the universities to foster competent graduate students. POSTECH was additionally selected as the operating university of this initiative last September.

POSTECH AI Graduate School plans to offer degree programs to educate students with its AI teaching experiences, know-hows, and optimal infrastructure. One of its strengths is young and eminent faculty members. So far, eleven faculty members are affiliated with the School and the number of faculty will be increased up to 26 by 2023.

While having received much attention and preparing for the launch of the programs, the AI Graduate School had its first admission of students last December and showed high selectivity. The acceptance rate was 18.5% for Masters, PhD and Masters-PhD Integrated degree programs. Among the three, MS-PhD Integrated program was the most selective with its acceptance rate of 9%.

POSTECH AI Graduate School has gone through a rigorous process for opening the academic programs and will open in March. Looking forward to great synergies among faculty, newly admitted students, and academic infrastructure, we interviewed the faculty who will lead the AI Graduate School and talked about their visions and plans.

◆ ‘AI + X’, one of the skills to develop, focusing on integration and collaboration


There was a boy who watched animated movies and admired artificial intelligence and robots. Later, he became the most prominent researcher in the field of computer vision and sketch visual recognition. This is Professor Suha Kwak of the AI Graduate School.

Professor Kwak completed his B.S. and PhD in computer science at POSTECH and worked as a post-doctoral researcher at École Normale Supérierue and Inria Paris in France. Shortly before he came to POSTECH, he was a faculty at Daegu Gyeongbuk Institute of Science and Technology (DGIST). He joined POSTECH in February of 2018.

His major interest of research is in computer vision. AI recognizes visual information such that it distinguishes objects shown on a display or divides areas. His research on weakly-supervised learning and visual recognition of AI with minimal human supervision is a world-class accomplishment.


“Before 2010, most of the research was conducted on crime analysis and surveillance that were often shown in movies. Recently, it has been expanded to wide ranges or areas such as automated vehicles, medical images, face-changing camera applications and more,” said Prof. Kwak.

He further said, “AI technology has been developing fast and can be found in our daily life, indicating various possibilities. The academic curriculum of the AI Graduate School at POSTECH is designed to best fit to this trend of technology and is expected to cultivate expertise in the field of AI.”

Prof. Kwak gave advice to his prospective students and said, “Personally, I regret that I could not study mathematics more in depth. To study AI, mathematical skills are basics and essential. This does not only apply to incoming students but to all students who wish to study AI and its related field. I recommend to build strong mathematical foundations.”


He also advised that the ability to collaborate with others is another essential skill needed. “In these days, it is difficult to accomplish competitive results alone. It is required to collaborate with others. One of the competencies that the Graduate School emphasizes is AI + X, the ability to integrate with various fields.”

As an example of this, Prof. Kwak and his research team are having regular meetings with Prof. Sanguk Kim and his research fellows at the Department of Life Sciences. In addition to this initiative, he is also exchanging research ideas with other groups of faculty who have similar research interests and discuss their possible collaborations and developing plans.

“Quantitative accomplishments such as papers, symposiums are important, however, it is my goal to conduct creative research that will lead the World. And I will do my best to pass this on to my research fellows and students,” said Prof. Kwak.

◆ Educating AI with What We Have Learned So Far


Prof. Jungseul Ok joined POSTECH in August of last year and it has been a little over six months. He is still unfamiliar with the University but showed his anticipation in the AI Graduate School in his interview.

Prof. Ok majored in electrical and electronic engineering at the Korea Advanced Institute of Technology (KAIST) and received his PhD in 2016 from the same institution. Before joining POSTECH, he worked as a postdoctoral researcher at KTH Royal Institute of Technology in Sweden and the University of Washington in the U.S.

Prof. Ok is focusing on machine learning and reinforcement learning. Especially, he is researching on the functional limitation of AI with minimal information programmed. The AI learning to perform tasks with no information or instruction gained much attention before, yet, in reality, there are optimal algorithms including basic data and destruction. In other words, he is researching how AI can accomplish optimal results with minimum information provided.


Based on this research, Prof. Ok was invited to the most prominent conference in AI, the Conference on Neural Information Processing Systems in 2018 and presented his paper as the first Korean author of the paper. At that time, only 30 speakers were selected for presentation among 4,856 researchers, which is about 0.6%. Also, he was an only Korean speaker.

Prof. Ok explained his understandings of AI as the expansion of problem-solving in various fields of agriculture, livestock industry, manufacturing industry and more. He said, “The AI Graduate School means that AI is considered as one of the academic disciplines and programs and it is very meaningful in that sense. The School will be an educational venue to learn the conventional AI theories and research and examine rapidly changing technology trend.”


As an essential element to realize this goal, he emphasized the competencies of the admitted students. “We have high expectations for the new incoming students, who are admitted to the highly selective programs at the AI Graduate School. Also, young faculty members have distinguished research competence and flexible mindsets that can incorporate with the fast-changing trend of AI,” Prof Ok said.

Research collaboration is also another advantage of the School. He explained, “Students in both undergraduate and graduate programs who have the same research motivation have gathered together and created joint research laboratories to work together. They are sharing research papers, setting research directions, suggesting new methods, and other activities that bring them synergies.”

Lastly, he advised students and said, “I think our research is about teaching machines how to learn. As the students themselves are already educational experts with over ten years of learning experiences since elementary school, I encourage them to discuss with each other and make a thriving school together.”

◆Becoming a Global Educational Venue for Fostering AI Skilled Talents


POSTECH AI Graduate School is planning to foster domestic students with essential AI skills beginning from this upcoming semester. The chief professor of the School, Young-joo Suh said, “In 2020, we plan to declare the opening of the AI Graduate School and to build a firm foundation by integrating with the conventional graduate schools at the same time. In addition, we have POSTECH’s academic, research, and academia-industry cooperation infrastructures and will foster top-notch students with AI skills.”

The School is also planning to continue hiring more faculty members and to establish an educational environment. Prof. Suh continued explaining, “It is important to secure AI experts with high degrees of engagement and investment in the field of AI. Last year, we invited six world-class scholars to our Graduate School as faculty and we plan to secure more world’s distinguished professors.”

The School has a plan to support students with global leadership empowerment. It will establish partnerships with Stanford University, Carnegie Mellon University, University of Michigan, and the University of Illinois at Urbana-Champaign and send students to these institutions on student exchange programs.

Also, students will have an opportunity to experience internships at global partner institutions such as NASA, Inria, Microsoft, Google, NVIDIA, Intel, and more. Plans for site visits to global AI corporations in Silicon Valley and financial supports for participating in international symposiums are underway.

Lastly, Prof. Suh shared his visions, “We are going to propose a role model to AI graduate schools in Korea with the most exemplary administration of the school. We will become a world’s prominent academic, research, academia-industry cooperative institution, taking the lead in the field of AI.”