学术报告一

报告题目：Learning theory for distributed learning

报告人：林绍波教授（研究员）（温州大学香港城市大学数学系）

报告时间：2019年4月30日（星期二）15：30-16：30

报告地点：数学与统计学院学术报告厅（25教14楼）

参加人员：教师、研究生、本科生

摘要：We derive optimal learning rates for distribute kernel ridge regression, provided the number of partitions is not so large. We also present some feasible strategies to improve the learning performance of distributed kernel ridge regression via enlarging the number of partitions. At last, we provide adaptive parametric selection scheme to fulfill the optimal learning performance for distributed learning.

报告人简介：林绍波，温州大学教授，香港城市大学数学系研究员。主要从事大数据算法设计与分析、分布式学习理论及深度学习理论的研究。在JMLR, IEEE TPAMI, ACHA, IEEE TSP, IEEE TNNLS, Constructive Approximation等国际著名期刊以第一作者或者通讯作者发表论文50余篇。主要贡献有：基于采样与算子理论，建立了分布式算法的理论基础；建立半监督分布式学习理论；提出了适定特征学习，大大减少了核学习的计算负担等。

学术报告二

报告题目：A Fast and Practical Randomized Method for Low-Rank Tensor Approximations

报告人：王尧副教授（西安交通大学统计学与大数据研究中心）

报告时间：2019年4月30日（星期二）16：30-17：30

报告地点：数学与统计学院学术报告厅（25教14楼）

参加人员：教师、研究生、本科生

摘要：Low-rank tensor approximations have gained much attention in dealing with real-world applications such as dynamic medical image processing and multi-channel video analysis, because of their efficiency in exploiting the intrinsic structure of the data with limited parameters. Unfortunately, the popular tensor decompositions that can get efficient low rank approximations, namely Tucker decomposition and tensor Singular Value Decomposition (t-SVD), for computing many SVDs are prohibitively computationally expensive in general, which obviously limits their use in“big data" environments. To remedy such issue, in this work, we present a randomized URV decomposition for producing fast and efficient low rank tensor approximations with theoretical guarantees by using Tucker decomposition and t-SVD. To be more precise, our method incorporates a strong rank-revealing QR decomposition that can make the computations of Tucker decomposition and t-SVD to be more stable. We then justify the effectiveness of the obtained low rank tensor approximations through a series of synthetic data experiments and several real-world applications. The extensive experimental results demonstrate the superior performance of our procedures over the existing methods in terms of both robustness and computational speed.

报告人简介：王尧，男，西安交通大学应用数学系与美国佐治亚理工学院工业与系统工程系联合培养博士，西安交通大学管理科学与工程博士后。现为西安交通大学统计学与大数据研究中心副教授。主要研究方向为机器学习与人工智能方法在图像视频数据分析、知识图谱、精准医疗以及推荐系统方面的应用。他已在IEEE TIP, IEEE TNNLS, IEEE TCYB, IEEE TSP, ICML, ICCV, CVPR等国际一流期刊与顶级学术会议上发表论文三十余篇，相关研究成果获陕西省科学技术一等奖和重庆市自然科学三等奖。曾为CVPR, ACCV等多个国际学术会议的PC。