【学术视频】第九届量子多体系计算研讨会 | 中国人民大学高泽峰

更多精彩视频登陆网站

www.koushare.com

图 | 高泽峰

题   目：Compressing deep neural networks by matrix product operators

报告人：高泽峰

单   位：中国人民大学

时   间：2019-04-21

地   点：中国人民大学

扫码观看精彩报告视频

报告提纲

• MPO can be used to represent both fully-connected and convolutional layers.

• MPO is more efficient in representing a fully-connected layer.

• MPO can reduce the cost in decoding each "image" in a dataset, and by combining with the MPO representation of the linear transformation matrices, can further compress deep neural networks and enhance its prediction power.

会议简介

This workshop focus on recent advances in the field of quantum many-body computation, including Monte Carlo methods, renormalization group and tensor network methods, machine learning, etc. It aims to provide young researchers and graduate students with introductory and updated research information, promote exchange of ideas and in-depth discussions, and enhance the interactions of researchers working in relevant areas.

Organizer:  Department of Physics, Renmin University of China

—— ——往期精彩回顾—— ——

END

扫描二维码，关注微信公众号

戳这里，观看精彩视频哟！