Title: | CMIV and ICTS Joint Colloquium (Lecture 1): Fundamentals of Data Assimilation I -- Ensemble Based Methods |
Speaker: | Prof. Adrian Sandu, Department of Computer Science, Virginia Tech |
Time/Place: | 10:30 - 11:30 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | This talk introduces fundamental concepts of data assimilation: errors in the data, errors in the model, errors in the prior, Bayesian inference, Kalman filter. We introduce Monte Carlo based filters that are able to cope with nonlinear dynamics: ensemble Kalman filter, ensemble transform filters, square root filters, and particle filters. We discuss important aspects like localization and inflation. |
Title: | CMIV and ICTS Joint Colloquium (Lecture 2): Fundamentals of Data Assimilation II -- Variational Methods |
Speaker: | Prof. Adrian Sandu, Department of Computer Science, Virginia Tech |
Time/Place: | 14:30 - 15:30 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | We review the variational approach to data assimilation to obtain maximum likelihood estimators. Techniques discussed include 3D-Var and 4D-Var. Aspects of the construction of adjoint models and optimization will be presented. Solution approaches like incremental 4D-Var, and the inclusion of model errors via weak constraint 4D-Var, will also be discussed. |
Title: | Fast Computation of Wave Propagation in Acoustical Waveguides Terminated by a PML |
Speaker: | Prof. ZHU Jianxin, Department of Mathematics, Zhejiang University, Hangzhou, China |
Time/Place: | 11:00 - 12:00 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | When an open acoustical waveguide is bounded by a perfectly matched layer (PML). The wave propagation equation – Helmholtz equation is changed to a complex PDE. However, the eigenfunctions of the PDE lost the orthogonal property, and it leads to the transformation difficulty of local bases by use of the eigenmode expansion method. In this talk, an adjoint eigenoperator is constructed, and its eigenfunctions are crossly orthogonal with the original ones. Applying this treatment to the computation of wave propagation, numerical simulations show that the computational efficiency is raised. |
Title: | Trace Pursuit: A general approach for model free variable screening and selection |
Speaker: | Dr. YU Zhou , School of Finance and Statistics, East China Normal University, China |
Time/Place: | 11:00 - 12:00 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | We propose trace pursuit for model-free variable selection under the sufficient dimension reduction paradigm. Two distinct algorithms are proposed: stepwise trace pursuit and forward trace pursuit, both of which can be combined with many existing sufficient dimension reduction methods. Stepwise trace pursuit achieves selection consistency with fixed p, and is readily applicable in the challenging p > n setting. Forward trace pursuit can serve as an initial screening step to speed up the computation in the case of ultrahigh dimensionality. The screening consistency property of forward trace pursuit is established. Finite sample performances of trace pursuit and other model-free variable selection methods are compared through numerical studies. |
Title: | Generalized Canonical Correlation Analysis and Its Application to Blind Source Separation |
Speaker: | Dr. LIU Wei, Department of Electronic and Electrical Engineering, University of Sheffield, United Kingdom |
Time/Place: | 14:30 - 15:30 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | Blind source separation (BSS) is one of the most important and established research topics in signal processing and many algorithms have been proposed based on different statistical properties of the source signals. For second-order statistics (SOS) based methods, canonical correlation analysis (CCA) has been proved to be an effective solution to the problem. In this talk, the CCA approach will be generalized to accommodate the case with added white noise and it is then applied to the BSS problem for noisy mixtures. In this approach, the noise component is assumed to be spatially and temporally white, but the variance information of noise is not required. An adaptive blind source extraction algorithm is derived based on this idea and a further extension is proposed by employing a dual-linear predictor structure for blind source extraction (BSE). |
Title: | A Concave Security Market Line |
Speaker: | Prof. Thierry Post, College of Administrative Sciences & Economics, Koç University, Turkey |
Time/Place: | 11:00 - 12:00 FSC1217, Fong Shu Chuen Library, HSH Campus, Hong Kong Baptist University |
Abstract: | We provide theoretical and empirical arguments in favor of a diminishing marginal premium for market risk. In capital market equilibrium with binding portfolio restrictions, investors with different risk aversion levels generally hold different sets of risky securities. Whereas the traditional linear relation breaks down, equilibrium can be described or approximated by a concave relation between expected return and market beta, and a concave relationship between market alpha and market beta. An empirical analysis of U.S. stock market data confirms the existence of a significant concave cross-sectional relation between average return and estimated market beta. We estimate that the market risk premium is at least five to six percent per annum, substantially above traditional estimates. A practical implication for active portfolio managers is that the alpha of "betting against beta" strategies seems dominated by the medium-minus-high-beta spread rather than the low minus-medium-beta spread. The success of such strategies thus largely depends on under-weighting or short selling high-beta stocks. |
We organize conferences and workshops every year. Hope we can see you in future.
Learn MoreProf. M. Cheng, Dr. Y. S. Hon, Dr. K. F. Lam, Prof. L. Ling, Dr. T. Tong and Prof. L. Zhu have been awarded research grants by Hong Kong Research Grant Council (RGC) — congratulations!
Learn MoreFollow HKBU Math