Topics on sparse identification of nonlinear dynamics(SINDy) theory and application

Prerequisite

Calculus, Mathematics Statistics

Abstract

Sparse Identification of Nonlinear Dynamics (SINDy) is a machine learning method proposed by Steven L. Brunton group to identify the form of differential equations. SINDy method has been widely used in various fields, such as real-time prediction of aeroelastic model in aerospace field, inference of gene control network in biochemistry field. At the same time, some researchers also give theoretical derivation of the convergence of sparse regression algorithm. This course will mainly introduce the theory and application of SINDy. In addition, classical machine learning methods, such as linear regression, nonlinear regression, model selection, feature extraction, k-means clustering, support vector machines, multilayer neural networks and activation functions, will be introduced.

Lecturer Intro.

杨武岳，毕业于清华大学，理学博士。从事生物数学、机器学习理论及其应用等研究。