报告摘要：The pathwise coordinate optimization is one of the most important computational frameworks for solving high dimensional convex and nonconvex sparse learning problems. The pathwise coordinate optimization differs from the classical block coordinate descent algorithms in two salient features: warm start initialization and active set identification. These two features grant superior empirical performance, but at the same time pose significant challenge to theoretical analysis. To tackle this long lasting problem, we develop a new theory showing that these two features play pivotal roles in guaranteeing the optimal statistical and computational performance of the pathwise coordinate optimization. In particular, our analysis provides new theoretical insights on the existing pathwise coordinate optimization framework and indicates its possible theoretical drawbacks. Based on the obtained insights, we modify the existing pathwise coordinate optimization framework and propose a new algorithm which guarantees to converge linearly to a unique sparse local optimum with good statistical properties (e.g. minimax optimality and oracle properties). This is the first result establishing the computational and statistical properties of the pathwise coordinate optimization framework in high dimensions. Thorough numerical experiments are provided to support our theory.
报告人简介：Tuo Zhao is a four year PhD student in Department of Computer Science at Johns Hopkins University. He received his B.S. and M.S. in Computer Science from Harbin Institute of Technology, China. He received his second M.S. in Applied Math from University of Minnesota. He is also affiliated with Statistics Laboratory at Princeton University. His research focuses on large-scale nonparametric learning and its applications to high throughput genomics and neuroimaging. He has published 14 papers on top journals and conferences, and developed several popular open-source software packages for high dimensional sparse modeling. He was the core member of the JHU team winning the INDI ADHD 200 global competition on fMRI imaging-based diagnosis classification in 2011. He is also active in adaptive clinical trial studies and gene expression network analysis. He co-authored a Nature paper on the exonic de novo mutations in autism spectrum disorders in 2012.
文章信息：Transferring Rich Feature Hierarchies for Robust Visual Tracking, 2015. [arxiv]
报告摘要：Convolutional Neural Networks (CNNs) have demonstrated its great performance in various vision tasks, such as image classification and object detection. However, there are still some areas that are untouched, such as visual tracking. We believe that the biggest bottleneck of applying CNN for visual tracking is lack of training data. The power of CNN usually relies on huge (possible millions) training data, however in visual tracking we only have one labeled sample in the first frame. In this paper, we address this issue by transferring rich feature hierarchies from an offline pretrained CNN into online tracking. In online tracking, the CNN is also finetuned to adapt to the appearance of the tracked target specified in the first frame of the video. We evaluate our proposed tracker on the open benchmark and a non-rigid object tracking dataset. Our tracker demonstrates substantial improvements over the other state-of-the-art trackers.
报告人简介：I am currently the final year PhD candidate in CSE department, HongKong University of Science and Technology. My supervisor is Prof. Dit-Yan Yeung. Before that, I got my BS degree from Zhejiang University, 2011 under the supervision of Prof. Zhihua Zhang. My research interest focuses on applying statistical computational model to real problems in computer vision and data mining. Currently, I mainly work on sparse representation, matrix factorization, deep learning. Especially I am interested in the area of visual tracking, object detection, image classification and recommender system.