当前位置: 网站首页 >> 员工动态 >> 正文
2023.07.05,曹延昭,教授,美国奥本大学,On a stochastic deep neural network


报告题目:On a stochastic deep neural network

报告人:曹延昭  教授 (美国奥本大学)

报告人简介: 曹延昭教授,1983年毕业于吉林大学数学系,1996年获弗吉尼亚理工学院数学博士学位,现任美国奥本大学数学与统计学系 Don Logan endowed chair in mathematics,应用与工业数学学会(SIAM)美国东南地区分会主席。主要从事偏微分方程和积分方程数值解法、随机偏微分方程数值解、非线性滤波、不确定性量化等领域的研究,部分研究成果发表在《SIAM J. Numer. Anal.》、《Numer. Math.》、《Math. Comp.》、《IMA J. Numer. Anal.》等计算数学顶级期刊上。现担任包括计算数学国际首要期刊《SIAM J. Numer. Anal. 》等学术期刊编委,研究课题得到美国国家自然基金会以及美国能源部的资助。

摘要 This presentation provides a mathematical introduction to deep learning, In particular, I will relate the deep neural network to traditional mathematical tools such as the Fourier series. This will be followed by a discussion on recent advancements in uncertainty quantification (UQ) of deep learning models. On this topic, I will present convergence results and share the findings from numerical experiments conducted within our framework. These results contribute to a better understanding of uncertainty quantification in deep learning and shed light on the performance and behavior of the proposed UQ methodology.

时间:202375日(周三)上09:00-10:00

腾讯会议号:101 150 437

员工观看教室:沙教一 201

欢迎大家参加!