Zongben Xu, Hai Zhang, Yao Wang, Xiangyu Chang. L1/2 regularizer. Science in China Series F: Information Sciences. 2009 .PDF Bin Zou, Louqing Li and Zongben Xu, The Generalization Performance of ERm Algorithm with Strongly Mixing Observations, Machine Leraning. 25(2): 188-200, 2009 PDF Ruihu Li, Zongben Xu and Xuejun Zhao, On The Classification of Binary Optimal Self-Orthogonal Codes, IEEE TRANSACTIONS ON INFORMATION THEORY, 54(2008), 3778-3782.   PDF Ruihu Li, Zongben Xu and Xueliang Li, Standard Forms of Stabilizer and Normalizer Matrices for Additive Quantum Codes, IEEE TRANSACTIONS ON INFORMATION THEORY, 54(2008), 3775-3778. PDF Deyu Meng, Yee Leung, Tung Fung and Zongben Xu, Nonlinear Dimensionality Reduction of Data Lying on the Multicluster Manifold, IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETIS—PART B: CYBERNETICS, 38(2008),1111-1122.PDF Initiated and sponsored by 973 Project “Theories and Techniques for Unstructured Information Processing Based on Visual Recognition, the workshop on “Visual Cognition: Principles, Modeling and Computation” was held from April 7 -11, 2008 in Xi’an Jiaotong University. 120 participants from Xi’an Jiaotong University, Tsinghua University and Shanghai Institute for Biological Sciences attended the workshop. Prof. Xu and his team won the State Natural Science Award, 2007 (Second Class). Premier Wen Jiabao  was shaking hands with him during the award conference held on January 8th, 2008 in the Great Hall of People, Beijing.  Prof. Zongben Xu was conferred the CSIAM Su Buchin Applied Mathematics Prize on the Fifth Delegation Meeting of Chinese Society for Industrial and Applied Mathematics and  The Second Award Conference of the CSIAM Su Buchin Applied Mathematics Prize, which was held in Zhengzhou on August 27，2008.

Machine Learning Group The research  mainly focuses on development of learning theory and methods for highly complex , large scale data, featured as cross-disciplinary point of view . The main interests include: statistical learning theory,  algorithmic stability ,  learning algorithms(neural networks,  SVMs,  boosting,  regularization,  etc.), variable selection and sparisity.   Vision Computing Group This group focuses on the mathematical analysis and modeling of visual principle. The main emphasis is on establishment of new computational model , and provision of novel solutions to fundamental  problems in data mining and computer vision in virtue of the new established  computational models.   High Dimentional Data Analysis Group The group mainly studies nonlinear dimensionality  reduction,  intrinsic  dimension estimation and  feature extraction theory and method of high dimensional and large scale data set. The  focus is on  validity and applicability of manifold learning algorithms in very complicated  data sets ,  fast  and  incremental manifold learning, etc. Data Mining Application Group   The group is mainly involved in application of  data mining technology in  financial  engineering , telecommunication and flow industry (especially in  iron and steel  production processes). Theories and Techniques for Unstructured Information Processing Based on Visual Recognition, National Basic Research Program of China (973 Program) (Chief Scientist) Model selection strategies for support vector regression, funded by the National Natural Science Foundation of China (principle Investigator)  Data mining technologies for high complex databases with application to management science, funded by the National Natural Science Foundation (Key Project) (Principle investigator)