Creating innovative bio-convergent technologies for better human life

이상완
homepage
Center for Neuroscience-inspired Artificial Intelligence
Sang Wan Lee, Ph. D

Director

인간의 두뇌가 가지고 있는 무한한 잠재력을 기술의 영역으로 풀어내고 이를 인공지능으로 이식하는 신경과학-인공지능 융합연구는 현재 인공지능의 수준을 한단계 높이는 출발점이며, 더 나아가 인간과 인공지능이 함께 진화해 나갈 수 있는 미래사회의 밑그림을 그려가는 과정입니다. 본 연구센터는 이러한 차세대 뇌기반 인공지능 연구를 선도하기 위해 끊임없이 고민하고 노력할 예정입니다.

 

KAIST Center for Neuroscience-inspired AI (CNAI) pursues a multidisciplinary approach to understanding the computational and neural basis of intelligence, in order to achieve neuroscience-inspired solutions for next-generation artificial intelligence.

 

Our challenge: Limitations to engineering intelligence without understanding intelligence.

 

Artificial intelligence (AI) technologies have evolved from data-driven types into a self-play type. This new class of algorithms has demonstrated an ability to succeed in various high-level tasks, even emerging as a general framework for decision-making in neuroscience and robotics. However, the brain is equipped with a multitude of functions that even the state-of-the-art AI does not yet possess.

 

Our approach: Advancing artificial intelligence using insights from cognitive and brain sciences.

 

The Center for Neuroscience-inspired AI (CNAI), located at KAIST, aims to understand the computational and neural basis of biological intelligence and to use that knowledge as a basis for developing advanced neuroscience-inspired AI technologies. Research at CNAI places particular emphasis on collaboration across various fields that have thus far existed largely independently, including developmental psychology, computational neuroscience, neural engineering, machine learning, and robotics.

 

Our goal: Development of neuroscience-inspired AI with human-like traits.

 

We believe that the implementation of AI that learns and processes information in the same way as humans will enhance the capability of AI to offer better solutions to various engineering challenges. Moreover, an important implication arising from the development of human-like AI is its ability to understand and compute human “suboptimal” flexible traits. This would provide an ideal setting in which humans interact intuitively and effectively with AI-powered tools. Our center is supported by the Ministry of Science and ICT, Republic of Korea.