Colloquium/Seminar

YearMonth
2017 Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Oct   Nov   Dec  
2016 Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Oct   Nov   Dec  
2015 Jan   Feb   Mar   Apr   May   Jun   Aug   Sep   Oct   Nov   Dec  
2014 Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Sep   Oct   Nov   Dec  
2013 Jan   Feb   Mar   Apr   May   Jun   Aug   Sep   Nov   Dec  
2012 Jan   Feb   Apr   May   Jun   Jul   Aug   Sep   Nov   Dec  
2011 Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Sep   Oct   Nov   Dec  
2010 Jan   Feb   Mar   Apr   May   Jun   Sep   Oct   Nov   Dec  
2009 Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Sep   Oct   Nov   Dec  
2008 Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Sep   Oct   Nov   Dec  
2007 Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Sep   Oct   Nov   Dec  
2006 Jan   Feb   Mar   Apr   May   Jun   Jul   Sep   Oct   Nov   Dec  
2005 Jan   Feb   Mar   Apr   May   Jun   Jul   Aug   Sep   Oct   Nov   Dec  
2004 Jan   Feb   Mar   Apr   May   Aug   Sep   Oct   Nov   Dec  

Event(s) on August 2016


  • Friday, 5th August, 2016

    Title: SPLBoost: An Improved Robust Boosting Algorithm Based on Self-paced Learning
    Speaker: Dr. WANG Yao, Department of Statistics, Xi’an Jiaotong University, China
    Time/Place: 14:00  -  15:00
    FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University
    Abstract: It is known that Boosting can be interpreted as a gradient descent technique to minimize an underlying loss function. Specifically, the underlying loss being minimized by the traditional AdaBoost is the exponential loss, which is proved to be very sensitive to random noise/outliers. Therefore, several Boosting algorithms, e.g., LogitBoost and SavageBoost, have been proposed to improve the robustness of AdaBoost by replacing the exponential loss with some robust loss functions. In this work, we present a new way to robustify AdaBoost, i.e., incorporating the robust learning ideas of Self-paced Learning (SPL) into Boosting framework. Specifically, we design a new robust Boosting algorithm based on SPL regime, i.e., SPLBoost, which can be easily implemented by slightly modifying off-the-shelf Boosting packages. Extensive experiments and a theoretical characterization are also carried out to illustrate the merits of the proposed SPLBoost.


  • Monday, 15th August, 2016

    Title: A Kernel Collocation Method by Using the Restricted Sobolev Kernels and L^2 Convergence Analysis for PDEs on Manifolds
    Speaker: Ms. CHEN Meng, Department of Mathematics, Hong Kong Baptist University, Hong Kong
    Time/Place: 14:30  -  15:30
    FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University
    Abstract: We apply the unsymmetric strong-form meshless collocation method, known as the Kansa method, to solving strongly second-order elliptic partial differential equations on some smooth, connected and closed manifolds. Based on the restricted Sobolev kernels and some assumptions on manifolds, the L^2 convergence theory is developed with dense requirements of collocation points for a least-squares problem. And besides continuous differential operators, we generalize implementations of discrete differential operators, which do not require derivatives of normal vectors or matrices to manifolds, by using the Kansa method. Lastly, numerical results are compared to verify convergence rates between the Kansa method and the symmetric RBF interpolation method, as well as numerical and theoretical collocation settings.


  • Monday, 15th August, 2016

    Title: An Adaptive Least Square Meshless Method for Cauchy Problem
    Speaker: Ms. LI SIQING, Department of Mathematics, Hong Kong Baptist University, Hong Kong
    Time/Place: 15:30  -  16:30
    FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University
    Abstract: An adaptive least square meshless method is proposed for solving the elliptic cauchy problem. The method use at most three step in computation. Tikhonov regularization naturally appear in the least square formulations and regularization parameter can be evaluated exactly. The convergency of numerical solution to the analytic one is proved based on the scatter data approximation theory in reproducing kernel Hilbert space. Numerical experiments in both two dimensional and three dimensional show the effective of the proposed method in estimate the unknown data from the noisy cauchy data, especially in the case when the noise level is large.


  • Monday, 15th August, 2016

    Title: New semiparametric regression method with applications to selection-biased sampling and missing data problems
    Speaker: Dr. DIAO Guoqing, Department of Statistics, George Mason University, USA
    Time/Place: 16:00  -  17:00
    FSC1111, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University
    Abstract: We propose a new method to estimate a regression function based on the semiparametric density ratio model, which can be viewed as a generalized linear model with a canonical link function and an unspecified baseline distribution function. Under this model, the distribution of the observed data retains the same structure in the presence of selection-biased sampling or when the predictors are missing at random. Particularly, in the latter case, the new method utilizes all the available information and does not need to specify the distribution of the predictors or the probability of observing the predictors. We establish large sample properties of the proposed regression estimators. Simulation studies demonstrate that the proposed estimators perform well in practical situations. Empirical data from the National Health and Nutrition Examination Survey are presented. This is joint work with Dr. Jing Qin.


  • Tuesday, 16th August, 2016

    Title: Three-Dimensional Wind Profile Prediction with Trinion-Valued Adaptive Algorithms
    Speaker: Dr. LIU Wei, Department of Electronic and Electrical Engineering, University of Sheffield, UK
    Time/Place: 11:00  -  11:30
    FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University
    Abstract: The problem of three-dimensional (3-D) wind profile prediction is addressed based a trinion wind model, which inherently reckons the coupling of the three perpendicular components of a wind field. The augmented trinion statistics are developed and employed to enhance the prediction performance due to its full exploitation of the second-order statistics. The proposed trinion domain processing can be regarded as a more compact version of the existing quaternion-valued approach, with a lower computational complexity. Simulations based on recorded wind data are provided to demonstrate the effectiveness of the proposed methods.


  • Tuesday, 16th August, 2016

    Title: Sparse Antenna Array Design for Directional Modulation
    Speaker: Dr. LIU Wei, Department of Electronic and Electrical Engineering, University of Sheffield, UK
    Time/Place: 11:30  -  12:00
    FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University
    Abstract: Directional modulation (DM) can be achieved based on uniform linear arrays (ULAs), where the maximum half wavelength spacing is needed to avoid spatial aliasing. To exploit the degrees of freedom (DOFs) in the spatial domain, sparse arrays can be employed for more effective DM design. In this work, the problem of antenna location optimisation for sparse arrays in the context of DM is addressed, where compressive sensing based formulations are proposed employing the group sparsity concept. Design examples are provided to verify the effectiveness of the proposed designs.


  • Friday, 19th August, 2016

    Title: A Penalized Multivariate Linear Mixed Effects Model for Integrative Proteo-genomics Analysis Based on iTRAQ Data
    Speaker: Dr. WANG Pei, Department of Genetics and Genomics Sciences, Institute of Genomics and multiscale Biology, Mount Sinai School of Medicine, USA
    Time/Place: 15:00  -  16:00
    FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University
    Abstract: Recent development in high throughput proteomics and genomics profiling makes it possible to study regulations of genetic factors on protein activities in a systematic manner. In this paper, we propose a new statistical method --- ProMAP --- a penalized multivariate linear mixed effects model for integrative proteo-genomic analysis. The motivation problem is to systematically characterize the regulatory relationships between proteins and DNA copy number alterations in breast tumor samples based on iTRAQ (isobaric tag for relative and absolute quantitation) data and SNP array data from CPTAC-TCGA studies. Because of the dynamic nature of iTRAQ technique and mass spectrometry instruments, data from iTRAQ experiments usually have severe batch effects, high percentages of missing and non-ignorable missing-data patterns. Thus, we utilize a linear mixed effects model to account for the batch structure and explicitly incorporate the batch-level abundance-dependent-missing-data mechanism of iTRAQ data in ProMAP. In addition, we employ a multivariate regression framework to characterize the multiple-to-multiple regulatory relationships between DNA copy number alterations and proteins. Moreover, we utilize proper statistical regularization to facilitate the detection of master genetic regulators, which affect the activities of many proteins and often play important roles in genetic regulatory networks. The performance of ProMAP is illustrated through extensive simulation studies. In the end, we apply ProMap to the CPTAC-TCGA breast cancer data sets, and identify novel regulatory relationships between DNA copy number alterations and protein expression profiles in breast cancer tumors.


  • Wednesday, 31st August, 2016

    Title: Solving Inverse Imaging Problems using Graph-Signal Smoothness Priors
    Speaker: Prof. Gene Cheung, Department of Electronic and Computer Engineering, Hong Kong University of Science and Technology, Hong Kong
    Time/Place: 16:00  -  17:00
    FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University
    Abstract: Inverse imaging problems like denoising, interpolation and bit-depth enhancement are inherently ill-posed, and signal priors are often used for regularization. A recent popular prior is the graph-signal smoothness prior: a desired image patch—interpreted as a graph-signal on an appropriately chosen graph—is assumed to be smooth with respect to the underlying graph. In this talk, I will first explain why such a signal prior is sensible from a graph signal processing (GSP) perspective. Then, I will describe a new graph-signal smoothness prior called LERaG based on left eigenvectors of the random walk graph Laplacian matrix, which has many desirable image filtering properties yet is computation-efficient. Finally, I will describe how LERaG can be used for soft decoding of JPEG images, resulting in state-of-the-art restored image quality.