Yilin Cheng
About
Projects
Design of long-context engineering algorithms for large language models
Build complex task datasets and design automatic context management algorithms to improve the model's reasoning ability in multimodal long sequence scenarios. (In progress)
Research on the theory and algorithms of data-driven large-scale optimization problems
The theory, application and design, development and analysis of data-driven large-scale optimization problems and their solution algorithms. Research on improving the semi-smooth Newton method to improve the efficiency of solving nonlinear semi-definite programming has been completed.
Research on Modern Nonlinear Methods for Stochastic Impulsive Differential Equations
Modern nonlinear methods are used to study the existence theory, periodic problems and related applications in finance and other fields of random impulse differential equations. Using the theory of topological degree and superposition degree, the existence, boundedness and boundary value properties of periodic solutions of second-order nonlinear random impulse differential equations are proved.
Education
PhD in Statistics
Fudan University
Bachelor of Mathematics
Hunan University