Speaker: Prof. Ning Chu (初宁博士)

Time: 12:10-13:10, 27  October 2023 (Friday) (Beijing time)

Venue: C305,Lijiao Building

Tencent Meeting ID：266-410-743

Abstract

Classical methods for inverse problems are mainly based on regularization theory. In particular those  which are based on optimization of a criterion with two parts: a data-model matching and a regularization term. Different choices for these two terms and great number of optimization algorithms have been proposed with great success. But there are still many limitations. Between them, the main one is accounting for the errors and uncertainty quantification.

Bayesian inference is the main tool to push much farther these limitations. Bayesian learning is the flexible tool for inference (inversion), learning (identification) and uncertainty quantification (UQ). However, the Bayesian computations can become very heavy computationally.

In this talk, an overview of these subjects is presented, particularly adapted for engineers. Few industry examples illustrate the main idea.

About Prof. Chu

初宁博士，教授级高工， 国际电气电子工程学会高级会员，浙江省声学学会副理事长；取得国防科大信息工程学士、法国巴黎萨克雷大学自动化硕士、信息科学博士；曾任瑞士洛桑联邦理工博士后，现任浙江上风高科首席研究员、副总工。从事贝叶斯深度学习，通风装备的绿色智能互联网+，在国际上率先研制成功的通风装备“工业芯肺系统”，于中央二套经济半小时首个专题报道、入选浙江省工业互联网平台建设名单、荣获工信部等国家四部委优秀智能制造场景案例。