报告人：雷娜
报告地点：腾讯会议
报告时间：2020年06月19日星期五10:00-11:00
邀请人：徐英祥
报告摘要：
Generative Adversarial Net (GAN) is a powerful model of machine learning, and becomes extremely successful recently. The generator and the discriminator in a GAN model competes each other and reaches the Nash equilibrium. GANs generate samples automatically, therefore reduce the requirements for large amount of training data. It can also model distributions from data samples. In spite of its popularity, GAN model lacks theoretic foundation. In this talk, we give a geometric interpretation to optimal mass transportation theory, explain the relation with the Monge-Ampere equation, and apply the theory for the GAN model. Based on this theoretic interpretation, we propose an Autoencoder-Optimal Transportation map (AE-OT) framework, which is partially transparent, and outperforms the state of the arts.
会议网址：https://meeting.tencent.com/s/aUo5G56Y6KFN
会议ID：475 674 840
会议密码：202006
主讲人简介：
雷娜，大连理工大学国际信息与软件学院书记，教授，博士生导师，几何计算与智能媒体技术研究所所长，北京成像技术高精尖创新中心研究员，纽约州立大学石溪分校计算机系访问教授；德克萨斯大学奥斯汀分校计算工程与科学研究所JTO research fellow；清华大学数学科学中心访问教授；。IEEE Transactions on Visualization and Computer Graphics，Computer Aided Geometric Design, Computer-Aided Design, Graphical Models等国际期刊审稿人, International Joint Conference of Artificial Intelligence, Geometric Modeling and Processing, Asian Conference on Design and Digital Engineering 等国际会议的 PC member. 研究方向为：应用现代微分几何和代数几何的理论与方法解决工程及医学领域的问题，主要聚焦于计算共形几何、计算拓扑、符号计算及其在人工智能、计算机图形学、几何建模和医学图像中的应用。主持国家自然科学基金重点项目、军科委创新特区项目等多项；多次受邀在国际、国内重要会议上做大会报告及会前课程。
学校主页