Welcome to Qinxin’s homepage!

I am Qinxin Yan, currently a Postdoc at University of Berkeley, in the group of Prof. Xin Guo.

I obtained my PhD in the Program in Applied and Computational Mathematics (PACM) at Princeton University. I was fortunate to be advised by Prof. H. Mete Soner.

My research develops asymptotic mathematical frameworks for large stochastic and strategic systems, combining control and game theory to analyze and design decision mechanisms in multi-agent settings. I blend stochastic simulation, mean-field theory, and graph-theoretic modeling with machine-learning–based algorithms to connect rigorous theory with applications in real-world problems, especially in finance and machine learning.

During 2023–2024, I was a visiting scholar with the Insurance Mathematics and Stochastic Finance Group (Group 3) in the Department of Mathematics at ETH Zurich. There, I worked with Prof. Josef Teichmann on machine learning algorithms for Mckean-Vlasov control, and with Prof. Beatrice Acciaio on the mean–field formulation of large neural networks.

My current projects include:

• Machine-learning-based methods for Mckean-Vlasov control problems and multi-agent games.
• Connections between Mckean-Vlasov control, Wasserstein gradient flows, and optimal transport.
• Applications of mean–field theory and Wasserstein gradient flows in large-scale neural networks.
• Viscosity solutions of Hamilton–Jacobi equations arising in Mckean-Vlasov control.
• Large stochastic particle systems with singular interactions on sparse graphs, with applications to systemic risk in financial networks.

Please explore my research to learn more.