□ DGIST (President Kunwoo Lee) has been selected for the new 2025 AI Star Fellowship project, hosted by the Ministry of Science and ICT Institute for Information & Communication Technology Planning & Evaluation, and will begin building a “human-centered symbiotic embodied artificial intelligence (AI) system.” This is a new, major project supported by the government to foster world-class AI specialists.
□ Over the course of six years (2025–2030), DGIST will receive a total of 11.5 billion won in research funding from the government for this project. The institute plans to develop a real-time embodied autonomous robot platform that can precisely understand human intentions and emotional states and actively adapt to environmental changes by studying bio-signals and behavioral data in various living and industrial settings.
□ DGIST will fuse key technologies, including ▲multimodal signal processing, ▲adaptive AI learning techniques, ▲optimized lightweight AI models, and ▲ neural processing unit-based computational structure designs. These technologies are expected to serve as the cornerstone for the development of intelligent robots that can coexist and interact naturally with humans.
□ Leading Korean companies such as LG Electronics, COGA Robotics, and FuriosaAI are participating in this project and will collaborate in various areas, including in the building of AI testbeds, data collection for learning and assessment, provision of experimental environments, and testing technologies. DGIST will establish a utilization model for empirical data and models, while also pursuing commercialization of research outcomes and potential industrial expansion.
□ DGIST will focus on advancing the decision-making of AI robots based on bio-signals and behavioral data, which have not been sufficiently covered in existing research. Through the application of unconscious reactions or emotional states, this technology is expected to play a key role in creating a next-generation “embodied AI system[1]” that can meticulously read and react to complex human intentions.
□ “Through this project, DGIST will secure its own unique physical AI technologies and foster world-class AI talent,” said Professor Park Kyung-joon of the Department of Electrical Engineering and Computer Science. “We will produce experts who will lead the physical AI era through differentiated technologies unique to DGIST that converge software, hardware, and platform.”
□ DGIST will continue to pioneer new paradigms in human–robot interactions by fostering emerging research talent, advancing data-based AI models, and closely collaborating with industries. It aims to heighten its competitiveness as a global university by establishing a strategic and sustainable research foundation that focuses on the three strategic areas of “physical AI and PI,” “digital twinning,” and the “quantum world.”
[1] Embodied AI: a physical AI (e.g., robot) that interacts with the real world and humans and has the ability to recognize its environment and react accordingly through sensors and actuators (mechanically moving parts).
