报告题目：The Evolution of Tensor Network Decompositions

报告时间：2023/03/22 14:30-15:30；

报告方式：#腾讯会议 664-451-265

摘要：Recently,  tensor network decompositions are emerging for capturing the intrinsic structures of multi-dimensional data, especially for high-order data. The topology design of tensor network decompositions is the key problem in the development of tensor network decompositions.  In this talk, I will first review the recent progress of tensor network decompositions. Then, I will discuss how to design fast algorithms for tensor network decompositions from a unified framework. Finally, we will also discuss the limitation and future possibilities of tensor network decompositions.

个人简介：赵熙乐，电子科技大学教授，入选电子科技大学百人计划和四川省学术和技术带头人后备人选。已第一或通讯作者在权威SIAM 系列期刊(SIIMS和SISC)和IEEE系列期刊(TIP、TNNLS、TCYB、TCI和TGRS)及计算机学会A类会议CVPR、AAAI和ACMMM等发表研究工作。研究成果获四川省科技进步一等奖两项(自然科学类、科技进步类)。