Mingyuan Zhou

 

I have moved to the University of Texas at Austin. My new homepage is here.

I'm a PhD student at Duke University and working as a research assistant in Professor Lawrence Carin's machine learning research group. My research lies at the intersection of Bayesian statistics and machine learning. I am interested in developing statistical theory and methods, hierarchical models, efficient Bayesian inference, and large-scale online and parallel learning algorithms. I am primarily focused on the development of nonparametric Bayesian priors, their hierarchical constructions, and latent variable models. Our research has led to novel nonparametric Bayesian latent variable models for both Gaussian and count data.

My recent research focuses on negative binomial process count and mixture modeling (including topic modeling). Please check our AISTATS 2012, ICML 2012 and NIPS 2012 publications and arXiv:1209.3442 for details.

Email: mingyuan.zhou@duke.edu; mz1@ee.duke.edu 
Phone: 919-660-5232 begin_of_the_skype_highlighting FREE 919-660-5232end_of_the_skype_highlighting
Mail Address: 131 Hudson Hall, Box 90291, Duke University, Durham, NC 27708
Curriculum Vitae     Google Scholar    VideoLectures

What's New

02/2013, I will join the University of Texas at Austin as an Assistant Professor of statistics in July 2013.

02/2013, Gibbs sampling Matlab code for "Lognormal and Gamma Mixed Negative Binomial Regression" is available for download.

10/2012, Gibbs sampling and Variational Bayes Matlab code for the negative binomial distribution is available for download.

We discover a Poisson-Logarithmic bivariate count distribution that can jointly model the counts of customers and tables:

                            

09/2012, "Negative Binomial Process Count and Mixture Modeling" was submitted, arXiv:1209.3442.

09/2012, "Augment-and-Conquer Negative Binomial Processes" was accepted to NIPS 2012 as a Spotlight Presentation.

04/2012, A paper on Bayesian Negative Binomial Regression was accepted to ICML 2012.

12/2011, A paper on Beta-Negative Binomial Process and Poisson Factor Analysis was accepted to AISTATS 2012.

Software

  1. Matlab codes and inference equations for "Non-Parametric Bayesian dictionary learning for sparse image representations" can be found  HERE
  2. BPFA Gray-scale and RGB image denoising code (last update 04/15/2010) can be found  HERE
  3. BPFA Gray-scale, RGB and Hyperspectral image inpainting & denoising code (04/15/2010 version, last update 02/16/2012) can be found  HERE
  4. Gibbs sampling Matlab code for the negative binomial distribution (10/12/2012 version) can be found  HERE
  5. Variational Bayes Matlab code for the negative binomial distribution (10/13/2012 version) can be found  HERE
  6. Gibbs sampling Matlab code for "Lognormal and Gamma Mixed Negative Binomial Regression" (02/16/2013 version) can be found  HERE

Selected Publications  [bibtex]  [Full List]

  1. Mingyuan Zhou and Lawrence Carin, "Negative Binomial Process Count and Mixture Modeling,"  arXiv:1209.3442, Sept. 2012.
  2. Gungor Polatkan, Mingyuan Zhou, Lawrence Carin, David Blei, and Ingrid Daubechies, "A Bayesian nonparametric approach to image super-resolution," arXiv:1209.5019, Sept. 2012.
  3. Q. Wu, D. Carlson, W. Lian, M. Zhou, C. R. Stoetzner, D. Kipke, D. Weber, J. Vogelstein, D. Dunson, and L. Carin, "Sorting Electrophysiological Data via Dictionary Learning & Mixture Modeling," submitted, Oct. 2012.
  4. Mingyuan Zhou and Lawrence Carin, "Augment-and-Conquer Negative Binomial Processes," to appear in Neural Information Processing Systems (NIPS2012), Lake Tahoe, NV, Dec. 2012. PDF Slides Poster (Spotlight oral presentation)
  5. Mingyuan Zhou, Lingbo Li, David Dunson and Lawrence Carin, "Lognormal and Gamma Mixed Negative Binomial Regression," International Conference on Machine Learning (ICML2012), Edinburgh, Scotland, Jun. 2012. PDF  Matlab Code  Appendix  Slides  Poster  Video  
  6. Mingyuan Zhou, Lauren Hannah, David Dunson and Lawrence Carin, "Beta-Negative Binomial Process and Poisson Factor Analysis," International Conference on Artificial Intelligence and Statistics (AISTATS2012), JMLR W&CP, 22:1462-1471, La Palma, Canary Islands, Spain, Apr. 2012.  PDF  Poster
  7. Xu Chen, Mingyuan Zhou and Lawrence Carin, "The Contextual Focused Topic Model," ACM SIGKDD Conf. Knowledge Discovery and Data Mining (KDD2012), Beijing, China, Aug. 2012.  PDF (Full Paper, Research Track)
  8. Lingbo Li, Xianxing Zhang, Mingyuan Zhou and Lawrence Carin, "Nested Dictionary Learning for Hierarchical Organization of Imagery and Text,''Conference on Uncertainty in Artificial Intelligence (UAI2012), Catalina Island, CA, Aug. 2012.  PDF
  9. Lingbo Li, Mingyuan Zhou, Guillermo Sapiro and Lawrence Carin, "On the Integration of Topic Modeling and Dictionary Learning," International Conference on Machine Learning (ICML2011), Bellevue, WA, Jun. 2011.  PDF
  10. Zhengming Xing, Mingyuan Zhou, Alexey Castrodad, Guillermo Sapiro and Lawrence Carin, "Dictionary Learning for Noisy and Incomplete HyperspectralImages," SIAM Journal on Imaging Sciences, Vol.5, pp. 33-56, Jan. 2012. PDF
  11. Mingyuan Zhou, Haojun Chen, John Paisley, Lu Ren, Lingbo Li, Zhengming Xing, David Dunson, Guillermo Sapiro and Lawrence Carin, "Nonparametric Bayesian Dictionary Learning for Analysis of Noisy and Incomplete Images," IEEE Trans. Image Processing, Vol. 21, pp. 130-144, Jan. 2012.  PDF  Matlab code and test results
  12. Mingyuan Zhou, Hongxia Yang, Guillermo Sapiro, David Dunson and Lawrence Carin, "Dependent Hierarchical Beta Process for Image Interpolation and Denoising," International Conference on Artificial Intelligence and Statistics (AISTATS2011), JMLR W&CP, 15:883-891, Ft. Lauderdale, FL, 2011.  PDF  Slides Video (Oral presentation)
  13. Mingyuan Zhou, Haojun Chen, John Paisley, Lu Ren, Guillermo Sapiro and Lawrence Carin, "Non-Parametric Bayesian dictionary learning for sparse image representations," Neural Information Processing Systems (NIPS), 2009.  PDF  Matlab code and Inference equations  Slides  Poster  Video (Oral presentation)
  14. Chengshi Zheng, Mingyuan Zhou and Xiaodong Li, "On the relationship of non-parametric methods for coherence function estimation," Signal Process., vol. 88, pp. 2863-2867, Nov. 2008.  PDF
  15. Mingyuan Zhou, Jialu Chen and Xiaodong Li, "A time/frequency-domain unified delayless partitioned block frequency-domain adaptive filter," IEEE Signal Process. Lett., vol. 14, pp. 976-979, Dec. 2007.  PDF
  16. Mingyuan Zhou and Xiaojun Qiu, "An error path delay compensated delayless subband adaptive filter architecture," Signal Process., vol. 87, pp. 2640-2648, Nov. 2007.  PDF

Other Publications

  1. Lingbo Li, Jorge Silva, Mingyuan Zhou and Lawrence Carin, "Online Bayesian Dictionary Learning for Large Datasets,"  to appear in International Conference on Acoustics, Speech and Signal Processing (ICASSP2012), Kyoto, Japan, Mar. 2012.  PDF
  2. Mingyuan Zhou, Hongxia Yang, Guillermo Sapiro, David Dunson and Lawrence Carin, "Landmark-Dependent Hierarchical Beta Process for Robust Sparse Factor Analysis," ICML2011 Structured Sparsity Workshop, Bellevue, WA,  Jun. 2011.  PDF
  3. Mingyuan Zhou, Hongxia Yang, Guillermo Sapiro, David Dunson and Lawrence Carin, "Covariate-Dependent Dictionary Learning and Sparse Coding,"  in Proc. International Conference on Acoustics, Speech and Signal Processing (ICASSP2011), Prague, Czech Republic, May 2011.
  4. Lingbo Li, Mingyuan Zhou, Eric Wang and Lawrence Carin, "Joint Dictionary Learning and Topic Modeling for Image Clustering," in Proc. International Conference on Acoustics, Speech and Signal Processing (ICASSP2011), Prague, Czech Republic, May 2011.
  5. Mingyuan Zhou, Chunping Wang, Minhua Chen, John Paisley, David Dunson and Lawrence Carin, "Nonparametric Bayesian Matrix Completion," in Proc. IEEE Sensor Array and Multichannel Signal Processing Workshop (SAM2010), Israel, Oct. 2010.  PDF
  6. John Paisley, Mingyuan Zhou, Guillermo Sapiro and Lawrence Carin, "Nonparametric Image Interpolation and Dictionary Learning Using Spatially-Dependent Dirichlet and Beta Process Priors," in Proc. International Conference on Image Processing (ICIP), Hong Kong, Sept. 2010. PDF Code
  7. Mingyuan Zhou, John Paisley and Lawrence Carin, "Nonparametric Learning of Dictionaries for Sparse Representation of Sensor Signals," in Proc. 3rd IEEE Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP2009), Aruba, 2009, pp. 237-240.  PDF  Slides 
  8. Mingyuan Zhou and Xiaodong Li, "A variable step-size for frequency-domain acoustic echo cancellation," in Proc. 2007 IEEE Workshop on Applications of Signal Processing to Audio and Acoustics (WASPAA'07), New Paltz, NY, Oct. 2007, pp. 303-306.  PDF

Professional Activities

Reviewer/Program Committee Member

- NIPS 2013, NIPS 2012,

- ICML 2013, ICML 2012

- IJCNN 2013, Special Session on Unsupervised Model-Based Learning: Bayesian regularization and Sparsity

- Journal of Machine Learning Research

- IEEE Transactions on Pattern Analysis and Machine Intelligence

- IEEE Transactions on Image Processing

- IEEE Transactions on Knowledge and Data Engineering

- IEEE Transactions on Circuits and Systems for Video Technology

- IEEE Geoscience and Remote Sensing Letters

- IEEE Signal Processing Letters

- SIAM Journal on Imaging Sciences

- Journal of Visual Communication and Image Representation 

- ACM Multimedia 2010

Invited/Contributed Talks:

- "Nonparametric Bayesian count and mixture modeling," McCombs School of Business and Division of Statistics & Scientific Computation, The University of Texas, Austin, TX, Jan. 2013.

- "Nonparametric Bayesian dictionary learning and count & mixture modeling," IBM T. J. Watson Research Center, Yorktown Heights, NY, Dec. 2012.

- "Augment-and-conquer negative binomial processes," NIPS 2012, Lake Tahoe, NV, Dec. 2012.

- "Nonparametric Bayesian count and mixture modeling," University of Southern California, hosted by Prof. Fei Sha, Los Angeles, CA, Oct. 2012.

- "Nonparametric Bayesian latent variable models," MERL - Mitsubishi Electric Research Laboratories, hosted by Dr. Dehong Liu, Cambridge, MA, July 2012.

- "Lognormal and gamma mixed negative binomial regression," ICML 2012, Edinburgh, Scotland, Jun. 2012.

- "Efficient Bayesian inference for the negative binomial distribution," Duke ECE Graduate Research Workshop, Durham, NC, Jan. 2012.
- "On the integration of topic modeling and dictionary learning," 8th Workshop on Bayesian Nonparametrics, June 2011, Veracruz, Mexico.
- "Covariate-dependent dictionary learning and sparse coding," ICASSP 2011, Prague, Czech Republic, May 2011.
- "Dependent hierarchical beta process for image interpolation and denoising," SampTA 2011, Singapore, May 2011.

- "Non-parametric Bayesian dictionary learning with landmark-dependent hierarchical beta Process," Learning Workshop, Ft. Lauderdale, FL, Apr. 2011

- "Dependent hierarchical beta process for image interpolation and denoising," AISTATS 2011, Ft. Lauderdale, FL, Apr. 2011.
- "Bayesian dictionary learning," Duke DISP Computational Imaging Seminar Series, Mar. 2011
- "Non-parametric Bayesian dictionary learning for sparse image representations," NIPS 2009, Dec. 2009, Vancouver, B.C., Canada.

- "Non-parametric Bayesian dictionary learning for sparse image representation and user rating matrix completion," UCLA, hosted by Prof. Stanley Osher, Los Angeles, CA, Nov. 2009.

Work Experience: Summer Intern at IBM T. J. Watson Research Center, Hawthorne, New York in 2010, working with Dr. Lexing Xie, Gang Hua and Apostol (Paul) Natsev.