Saitama University

Abstract：

Tensor networks have emerged as a powerful framework for addressing complex problems in fields ranging from quantum physics to machine learning. In this lecture, we introduce two cutting-edge methodologies within this domain: Tensor Cross Interpolation (TCI) and quantics tensor networks. TCI, a form of adaptive machine learning, excels at identifying hidden low-rank structures in data, enabling efficient representation and processing. On the other hand, the quantics representation is an independent concept that maps functions with exponentially varying length scales into tensor networks by discretizing their structures. By combining these two independent ideas, we unlock even greater potential, offering a robust framework for modeling high-dimensional data and solving complex problems across various disciplines. This lecture will be based on arXiv:2407.02454 and references therein.

Brief CV of Prof. Hiroshi Shinaoka:

Hiroshi Shinaoka is a computational physicist with a broad research scope, spanning from first-principles calculations of strongly correlated materials to the development and application of tensor network methods. Currently an Associate Professor at Saitama University, with a diverse academic background including international research experience at ETH Zurich. Actively involved in the development of open-source software and making significant contributions to computational physics and quantum material science.

报告时间：2025年2月12、13日、14日，上午10:00-11:30 

报告地点：M楼830会议室

主持人：王磊 研究员

联系人：傅琦（82649469）