Yao Xie

Change-point detection

Change-point detection is a fundamental problem in statistics and signal processing for applications such as threat detection and real-time monitoring. Given a sequence of high-dimensional observations, there can be a change-point in space and time, which alters the underlying distribution of the data. One would like to detect such a change as quickly as possible after it occurs.

I am interested in various change detection problems arising from modern applications, and aim to develop novel change-point detection methods that are computationally efficient, statistical powerful, as well as understand the fundamental limits, using techniques from statistics, optimization, and information theory.

Selected Publications

• Change detection via affine and quadratic detectors. (arXived version)
  Y. Cao, V. Guigues, A. Juditsky, A. Nemirovski, and Y. Xie.
• Multi-sensor slope change detection.
  Y. Cao, Y. Xie, and N. Gebraeel. Annals of Operations Research, pp. 1-27, April 2016.
• Sequential multi-sensor change-point detection.
  Y. Xie and D. Siegmund. Annals of Statistics, Vol. 41, No. 2, pp. 670-692, 2013.
• Change-point detection for high-dimensional time series with missing data.
  Y. Xie, J. Huang and R. Willett. IEEE Journal of Selected Topics in Signal Processing (J-STSP), vol. 7, no. 1, pp. 12-27, Feb. 2013.
• PMU based detection of imbalance in three-phase power systems.
  T. Routtenberg, Y. Xie, R. M. Willett and L. Tong. IEEE Transactions on Power Systems, Vol. 30, No. 4, pp. 1966 - 1976, July 2015. (arXiv)

Structured signal processing

Sparsity, low-rank, manifoldal and graphical structures are common geometric features of high-dimensional data. My interests are to develop efficient statistical methods that can exploit these structures to achieve more efficient signal estimation and detection algorithms, characterize their theoretical performance, and understand fundamental limits.

Selected Publications

• Poisson matrix recovery and completion.
  Y. Cao and Y. Xie. IEEE Transactions on Signal Processing, Vol. 64, No. 6, pp. 1609-1620, March 2016.
• Info-Greedy sequential adaptive compressed sensing.
  G. Braun, S. Pokutta, and Y. Xie. IEEE Journal Selected Topics in Signal Processing, Vol. 9, No. 4, pp. 601-611, June 2015. (Authors listed alphabetically.) (arXiv)
• On group coherence of frames.
  R. Calderbank, A. Thompson, and Y. Xie. Applied and Computational Harmonic Analysis, Vol. 38, pp. 50-71, 2015. (Authors listed alphabetically.) (arXiv)
• Reduced-dimension multiuser detection.
  Y. Xie, Y. C. Eldar and A. Goldsmith. IEEE Transactions on Information Theory, Vol. 59, No. 6, pp. 3858-3874, June 2013.

Network data analysis

Network is a fundamental model for interdependence between variables or physical quantities. Streaming data over networks have become ubiquitous nowadays, arising from sensors networks, social networks, power networks, internet-of-things and smart cities. My interest lies develop statistical framework for online learning and real-time inference of the network structures from noisy and incomplete observations.

Selected Publications

• Detecting weak changes in dynamic events over networks.
  S. Li, Y. Xie, M. Farajtabar, A. Verma, and L. Song. To appear. IEEE Transactions on Signal and Information Processing over Networks.
• Sequential change-point approach for community detection.
  D. Maragoni-Simonsen and Yao Xie. IEEE Signal Processing Letter, Vol. 22, No. 8, pp. 1035-1039, Aug. 2015. (arXiv)

Overview of my research prior to 2013: slides.