学术报告论坛报告一： Hyperspectral Image Processing: Orthogonal Constrained Minimization Models and
the Proximal Block Coordinate Descent Algorithms
报 告 人： 刘晓霞博士 香港理工大学
报告时间： 2024年12月23日（星期一）上午10:00-10:30
报告地点： 37号楼3A01报告厅
报告摘要：In this talk, we will present two orthogonal constrained minimization models for hyperspectral image processing. Each model incorporates a generalized tensor group sparsity measure and a (nonlocal) low-rank regularization, along with image priors such as nonlocal self-similarity or deep proximal denoisers. To solve the resulting nonconvex nonsmooth optimization problems with orthogonal constraints, we propose proximal block coordinate descent algorithms and demonstrate some convergence results. Experiments conducted on mixed noise removal and anomaly detection show that our proposed methods outperform state-of-the-art methods in terms of both numerical metrics and visual quality.
报告人介绍：Dr. Liu graduated from Sun Yat-sen University in 2012 with a Bachelor's degree and obtained her PhD in Mathematics from Syracuse University in 2018. After graduation, she worked at Shenzhen University as a postdoctoral fellow and later as an associate research fellow. In 2022, she joined the Hong Kong Polytechnic University as a research assistant professor. Her research interests include image inverse problems, low-rank representation, and sparse optimization.
论坛报告二: On the Kinetic Description of Objective Molecular Dynamics (OMD): multiscale modeling, numerics, data-driven applications
报 告 人: 綦昆仑博士 明尼苏达大学
报告时间： 2024年12月23日（星期一）上午10:30-11:00
报告地点： 37号楼3A01报告厅
报告摘要：In the first part of this talk, a multiscale modeling framework for objective molecular dynamics (OMD), a reduced molecular dynamics approach with inherent symmetries, will be presented. This hierarchical framework bridges OMD with statistical kinetic equations and macroscopic hydrodynamic models. In the kinetic regime, we identify two distinct interaction scalings, leading to either a mean-field-type or Boltzmann-type equation. At the macroscopic level, we derive reduced Euler and Navier-Stokes systems through a detailed asymptotic analysis. The second part of the talk introduces a fast spectral method for numerically solving the derived kinetic equations, supported by convergence analysis and numerical simulations to confirm its effectiveness. If time permits, I will discuss our recent progress in applying data-driven methodologies to kinetic theory.
报告人介绍：綦昆仑博士2017年本科毕业于华南理工大学数学学院，2021年7月博士毕业于香港城市大学，师从杨彤教授。2021年7月-2022年7月于香港中文大学进行博士后研究工作，并于2022年8月至今加入明尼苏达大学数学学院担任Dunham-Jackson助理教授(non-tenure track)。其研究方向集中于动理学理论(Kinetic Theory)，尤其是玻尔兹曼方程及其相关流体、材料模型的多尺度建模，数值模拟和误差分析，相关成果发表在SIAM J. Numer. Anal., SIAM J. Appl. Math., SIAM Multiscale Model. Simul., M3AS, JCP等期刊。
数学学院
2024年12月19日