As the modern, fast-paced machine learning theory tends to rely on a broad range of mathematical tools, there is a need for a communication nexus supporting leading, collaborative theoretical machine learning research of SNU's theory-oriented members. To this end, the SNU Theoretical AI research Group (STAG) supports regular and intimate interaction among its members who bring a broad theoretical background and coordinates with the graduate school of artificial intelligence. The key goals of STAG include probing new theoretical AI research directions, pursuing synergistic collaborative research, and teaching and advising students of the graduate school of artificial intelligence. We expect the activities of STAG to strengthen the theoretical machine learning research of SNU.
Just as Alan Turing set human intelligence as his prototype for intelligent behavior, the ultimate goal in the quest for artificial intelligence is the creation of machines that seamlessly incorporate human-like neurocognitive processes that underlie behavior (and to eventually surpass human levels of performance). However, human intelligence is a product of complex biological origins and, despite the huge advances in AI technology in the past few decades, the vast majority of existing AI algorithms designed to emulate human cognition, emotions, and behavior only do so at a superficial level, with little biological justification. It is perhaps due to this limitation that artificial intelligence still remains surprisingly inflexible and unadaptive in response to changing conditions. We believe that the next breakthrough in AI research will hinge on the construction and implementation of models that take into account how real brains process and organize environmental inputs and memory to guide behavior. Only then will we be able to engineer neuromorphic chips and artificial neural networks that allow truly cognitive AI to effectively navigate and interact with the environment, objects, and agents around them. The R.E.A.L. Intelligence Center defines four areas in which current AI systems can improve in light of natural intelligence: Recognition(R), Explainability(E), Adaptation(A), and Learning(L). By leveraging interdisciplinary research in engineering, computer science, neuroscience, and cognitive science, the center aims to make advances in these domains through principles taken directly from the scientific study of biological intelligence as instantiated in humans and other animal species.
Although AI is changing our lives, it is unknowable how internal AI processes work. Humans only ‘speculate’ on how the ‘hidden’ processes of AI work. The center studies human-AI interactions to better understand AI from a human perspective.
Just as architectural specifications are needed to construct a new building, we need computing specifications to make a new AI system. The university’s most renowned computing architects gathered at this center to develop ‘Specification of Computer Systems for Future AI’.
The Center for AI ELSI aims to examine the significant impact of artificial intelligence on our society and human values. As an interdisciplinary team of philosophers, legal scholars, psychologists, and social scientists, we comprehensively analyze the ethical, legal, and social issues raised by using artificial intelligence. Not only do we address the specific concerns in the medical, military, financial fields among others, but we comparatively review the various versions of AI ethics guidelines and suggest a general framework, which is to be tailored to our situations in Korea.
The process of drug development is long and complicated. The center started with an idea that AI might help in the process of drug development, especially during the stages of drug discovery and target validation.
SCAIHCare is a transdisciplinary research group that aims to develop, evaluate and disseminate innovative people-centered AI technologies promoting health, quality care, and wellbeing for all in aging societies in the era of global health security.
Center for Finance and Management AI focus on research for problems in finance and management domains. The center identify and solve the problems in finance and management by utilizing AI based approaches by interdisciplinary collaborations among researchers in various fields such as finance, management, industrial engineering, statistics and computer science. We aim to involve in and solve big problems which cannot be resolved without collaborations of many researchers with skills in the multiple fields, and contribute to the advancement of AI research in finance and management.
Linguistics researchers in different languages have gathered at this center aiming to publish a whitepaper on AI languages as a stepping stone for Korean language-based NLP research.
Transdisciplinary Center for Education in Seoul National University is a research center that aims to foster creative collaboration between AI and other areas of studies such as psychology, humanities, and pedagogy. It will help anyone in any AI-related projects find a solution to their own research or technological ideas. It will support and develop innovative research by applying AI technologies. There are three categories of projects in progress. First, ‘For the AI’ focuses on developing educational method for the current AI system. Second, ‘By the AI’ aims to improve educational contents by using AI, Third, ‘Of the AI’ emphasizes applying and utilizing digital transformation on new fields of studies.