Research
Our lab is working both on the study of fundamental theories and on the practical design and optimization of systems in the fields of communication, statistical inference, and machine learning. The followings are some topics on which our lab is currently working on.
Next generation communications
Currently, communication is at 5G, but the next generation will be 6G. In 6G, communication is expected to be significantly improved not only in terms of connectivity and latency, but also in providing intelligent communication services by utilizing a combination of terrestrial communication using base stations, aerial communication represented by UAVs, and space communication using low-orbit satellites. Our research lab is actively working on the design and optimization of systems for non-terrestrial network communication encompassing terrestrial, aerial, and space components.
Secure communications
It's important to make sure that no one listens our conversations without permission. Our research lab is working on how to keep communication safe so that information doesn't leak out. We're also looking into hiding communication, which means keeping the fact that communication is happening a secret. This is especially important in sensitive situations like military communications.
Differential privacy
Our data travels through various paths, like when you do computer searches, play games, or chat with friends on messengers. It's almost always being sent out to the outside world. During this process, unintended privacy leaks can become a serious problem. You might have just rated a movie, but your political preferences or religion could also be exposed. A technology that effectively and prominently protects against these privacy threats is called differential privacy. In our research lab, we're studying differential privacy techniques in a variety of data applications such as big data analysis, machine learning, and metaverse.
Quantum information theory
The 2022 Nobel Prize in Physics was awarded to three individuals for demonstrating quantum entanglement. The intriguing and counterintuitive properties of quantum phenomena like superposition and entanglement are expected to bring about revolutionary advancements in computing and communication. In our research lab, we are focused on studying the fundamental performance improvements that can be achieved by harnessing quantum entanglement for communication and data processing.
Federated learning
Federated learning is a machine learning paradigm where each user's device trains an AI model using their own data, and these individual AI models are then uploaded to a server. The server aggregates the AI models from each user to create a single global model, which is then sent back to the users. Users can further train this global model with their own data and upload it to the server again. By repeating this process, a good AI model, as if it had learned from all users' data, can be developed without the need for user data to leave their devices. Our research lab is working on the development of robust federated learning models against various challenging learning scenarios.