Yihong Wu joins ECE

ECE Assistant Professor Yihong Wu comes to the university from a postdoctoral fellowship with the statistics department at The Wharton School at the University of Pennsylvania.

Wu, who will be based in the Coordinated Science Laboratory, received the Marconi Society Paul Baran Young Scholar Award in 2011 and winner of the Best Student Paper Award, IEEE International Symposium Information Theory (ISIT) in 2011, received his PhD in electrical engineering with minors in mathematics, operational research, and financial engineering from Princeton University in September 2011. He received his BE in electrical engineering from Tsinghua University, Beijing, China, in July 2006.

Wu’s research combines information theory and statistics. One of the projects he is currently focusing on is sparse principal component analysis (sparse PCA). He is also interested in signal processing and communication problems.

“Principal component analysis is one of the major tools in dimension reduction in dealing with high-dimensional data,” said Wu. “We want to understand, ‘What is the optimal thing to do? What are the fundamental limits if we want to do PCA in the presence of sparsity in the data?’”

Wu’s research in modern statistics deals with large and complex data sets, and consequently with models containing a large number of parameters far exceeding the number of available samples.

“One example is genomics where the number of factors can be much larger than the number of subjects,” Wu said. “Conventional statistics usually operate in the regime of fixed dimension and large amounts of samples. In the presence of high-dimensionality, the fundamental question boils down to how to effectively take advantage of the underlying structure in order to perform optimal statistical inference efficiently.”

Wu’s research reflects his interests in information theory, mathematics, statistics, signal processing, communications, approximation theory, optimal transportation, and fractal geometry. He primarily focuses on information theory and statistics, which includes “applying information-theoretic tools to the modern high-dimensional statistics for applications like big data.”

“Big data usually refers to data of extremely large volume such as Web data, financial data, biological data, etc.,” Wu said. “Many data are naturally high-dimensional, which makes it very difficult to store, transmit, or perform inference on these data. For example, there are about 500,000 to 1 million observations along the human genome of just one individual. There are many facets of challenges for big data.”

Wu’s research focuses on the mathematical and statistical side. He uses tools that originate from information theory to determine the fundamental limit of statistical inference on high-dimensional data. That is, what is the optimal method to estimate, test and learn a high-dimensional object from data efficiently and adaptively?

Wu will teach courses in the area of information theory. He will be teaching a graduate-level course in the fall, and he hopes to develop a course that introduces the use of information theory for high-dimensional statistics.

Wu continues to collaborate with faculty at The Wharton School, and he hopes to work with faculty at the departments of ECE and Statistics at Illinois.

“I really like the excellent colleagues, outstanding students, and the collegial atmosphere at the Coordinated Science Lab and the ECE Department,” Wu said. “I am excited to be contributing to the innovative and multidisciplinary research at Illinois.”