[세미나] 문태섭 교수님

November 27, 2018

Introduction to machine (deep) learning and its application to PM2.5 estimation


문태섭 교수 (성균관대학교)

2018년 11월 27일 (화) 16:00

과학관 553호


Abstract

Introduction to machine (deep) learning and its application to PM2.5 estimation

In this talk, I will briefly introduce the main concepts of machine / deep learning within the supervised learning framework. Particularly, the so-called deep learning methods, which use hierarchical neural networks as function approximators, and their recent impacts on various application areas will be introduced. Furthermore, I will talk about some of the recent techniques about interpreting the prediction results of deep neural networks. Finally, I will give some application example of deep learning methods to PM2.5 estimation and show that our recent convolutional neural network-based approach can outperform traditional methods, e.g., random forest, in estimating US conterminous PM2.5 values.