Academic homepage of Dr. Yang Li

Dynamical systems, and qualitative theory for differential equations. We investigate the complex spatiotemporal pattern formation and evolution in such systems, as well as their potential relevance in biological systems. One of our current interests is the causal analysis based on an approach combining dynamical system theory and information theory.

Stochastic optimization. Currently, we are looking into the analysis of the performance of stochastic convex optimization algorithms (which adopt Monte-Carlo sampling, Nesterov's acceleration, etc).

 Sequence modeling based on recurrent neural networks and approaches of improving its performance. Currently,  our specific interest lies in the nonlinear time series analysis, including the prediction, classification, etc., of various data sets; specifically, we study reservoir computing, as a bridge that connect dynamical systems and time series analysis based on machine learning.