🎓About Me
Welcome to my personal website! I am Chenghan Xie (谢承翰), a Ph.D. student in the Operations Research group at Stanford University, advised by Professors Jose Blanchet and Renyuan Xu.
My research interests lie in Generative Models and AI Impact for Finance. My latest working paper is The AI Monoculture Paradox in Financial Markets, joint work with Jose Blanchet and Renyuan Xu on shared AI forecasts and information diversity.
Before joining Stanford, I obtained my Bachelor of Mathematics from Fudan University. I have conducted research under the supervision of Prof. Yinyu Ye and Prof. Ge. Additionally, I gained industrial research experience at ByteDance Seed and Shanghai AI Laboratory, working on diffusion model acceleration, multimodal/world models, and domain-specific LLM adaptation.
My long-term goal is to help define the right objectives for AI systems, so that their growing capabilities translate into improvements in human life, especially in financial markets. As LLMs become increasingly capable, I believe the central question is shifting from what they can do to what goals we should set for them. If you share similar interests, feel free to contact me via email.
- Generative Models and Diffusion Models
- AI Impact for Financial Markets
- Optimization and Large Language Model Inference
Ph.D. in Operations Research, 2024 - 2029 (expected)
Stanford University
B.Sc. in Mathematics, 2020 - 2024
Fudan University
Chenghan Xie’s current research focuses on generative models, diffusion models, and the economic impact of AI in financial markets.
My latest working paper, The AI Monoculture Paradox in Financial Markets, is joint work with Jose Blanchet and Renyuan Xu. We combine a Kyle-type market model with LLM-agent experiments to study how shared AI forecasts affect information diversity, trading profits, and adoption decisions. More accurate forecasts can lead to lower collective profits when traders rely on the same AI signal.
Read the paper on SSRN · Try Monoculture Lab, our interactive simulation platform