The role of communication is evolving beyond merely transmitting information to people, toward improving the cooperative efficiency of autonomous AI devices. The ultimate goal is to deliver essential information as accurately and quickly as possible, using limited resources to complete tasks. To contribute to this new paradigm, our lab focuses on developing integrated PHY and MAC algorithms for sensing, communication, and computation, along with communication protocols.

O-RAN is a redefined mobile communication system aimed at achieving openness, intelligence, interoperability, and virtualization. A key goal is to develop communication and network systems that can deliver AI services more flexibly and efficiently. In our lab, we focus on building O-RAN systems using software-defined radio and open-source software at the experimental level, and we are currently working on several national projects in this area.

With the rapid advancement of LLMs/LMMs, the versatility of machine learning technology has greatly increased. However, as model sizes grow, power consumption is rising exponentially. Additionally, as distributed learning—which involves collecting and sharing data across numerous networked nodes—advances, concerns over data reliability are becoming more prominent. In our lab, we focus on enhancing the energy efficiency of machine learning through techniques like network and communication-based task offloading, model lightweighting, and explainable AI. We are also developing machine learning technologies that ensure data reliability and are robust against adversarial attacks.