讲座人简介：

赵启斌，2009获得上海交通大学计算机系博士学位，目前在日本理化学研究所革新智能研究中心担任团队负责人，东京农工大学访问教授，研究兴趣包括机器学习、张量分解与张量网络，可信赖机器学习等，发表论文200余篇。担任机器学习会议NeurIPS、ICML等领域主席，以及期刊Neural Networks、Machine Learning的执行编辑。

讲座简介：

Modern ML methods have achieved the remarkable performance by dramatically increasing the DNN model size and the amount of high quality data samples. However, how to learn information from data efficiently and train a parameter efficient model become important in particular applications. Tensor Networks (TNs) have been increasingly investigated and applied to machine learning and signal processing, due to their advantages in handling large-scale and high-dimensional problems, model compression in DNNs, and efficient computations for learning algorithms. This talk aims to present some recent progresses of TNs technology applied to machine learning from perspectives of basic principle and algorithms, particularly in unsupervised learning, data completion, multi-model learning and various applications in deep learning modeling and adversarial robustness. Finally, we will also present potential research directions and new trends in this area.