CPS Computational Learning Laboratory
Research
The lab's ongoing research interests are as follows.Intelligent Sensing
Intelligent Sensing is a burgeoning research direction that combines advanced sensing technologies with intelligent algorithms to enable system or controller to perceive and understand the surroundings around them. Integrating intelligent sensing techniques into various applications holds great potentials. This multidisciplinary field encompasses various aspects such as computer vision, signal processing, statistics, machine learning, and so on. Intelligent sensing aims to enhance the ability of machines to acquire and interpret sensory information from their environment. To effectively model the sensory data, our lab adopts and develops machine learning and deep learning algorithms, which have demonstrated promising performances in real applications.
Complex System Analysis
Complex System Analysis focuses on studying and managing the behavior of complex systems composed of numerous interconnected components or agents. These systems exhibit intricate dynamics and emergent behaviors that cannot be fully understood or controlled by analyzing individual components in isolation. Complex systems often exhibit nonlinear dynamics. Our lab employs various mathematical and computational techniques. By employing mathematical, computational, and AI techniques, researchers aim to uncover the underlying dynamics, emergent behaviors, and vulnerabilities of these systems. The insights gained from this research can be applied to various domains, enabling more efficient, resilient, and adaptive systems.
Distributed Computing
Distributed Computing is a research direction that focuses on developing efficient algorithms and techniques for solving complex problems in distributed systems. In distributed computing, tasks are divided among multiple computational nodes or processors, which work collaboratively to solve a problem. Furthermore, the emergence of big data has further fueled the need for distributed computing. With the exponential growth of data, traditional centralized approaches struggle to process and analyze massive datasets efficiently. In the Internet of Things (IoT), distributed computing enables real-time processing and decision making on a large scale with distributed sensors and actuators. Our lab aims to solve large-scale distributed problems that cannot be tackled by traditional centralized approaches due to computational constraints.
Distributed Optimization
Distributed optimization algorithms leverage techniques such as decomposition, parallelism, and consensus-based methods to break down the problem into smaller sub-problems that can be solved independently by different nodes. Then distributed solutions are combined to obtain the overall global optimal solution. The developed distributed optimization algorithms can effectively distribute the workload among the nodes and balance the computational resources based on factors such as task requirements, communication costs, and available processing power. In addition to scalability and efficiency, distributed optimization also address issues related to fault tolerance and robustness in distributed systems. The system becomes more resilient to failures, as the failure of a single node or component does not lead to a complete system failure. Our lab aims to develop techniques to ensure the reliability and availability of distributed systems.