Fisher Decorator: Refining Flow Policy via A Local Transport Map
Conference on Neural Information Processing Systems (NeurIPS), 2026 · Poster
CV
Education
PhD, Nanyang Technological University, Singapore (Aug 2026–present)
College of Computing and Data Science (CCDS)
Supervisor: Prof. Yingzhen Li
Scholarship: NTU Research Scholarship (RSS)MSc, Applied Computational Science and Engineering, Imperial College London (2024-2025)
Department of Earth Sciences and Engineering | Distinction
Key Courses: Computational Mathematics (A*), Machine Learning (A*), Fluid Dynamics (A*), Deep Learning (A*)BSc, Applied Mathematics, Taiyuan University of Technology (2020-2024)
Department of Mathematics | GPA: 4.5/5.0 (2/40)
Key Courses: Mathematical Analysis (97), Advanced Algebra (98), Numerical Analysis (94)
Awards: Academic Outstanding Individual (2020-2024)
Publications
Scaling World-Model Reinforcement Learning Through Diffusion Policy Optimization
arXiv:2605.26282, 2026
On Stability and Robustness of Diffusion Posterior Sampling for Bayesian Inverse Problems
Conference on Neural Information Processing Systems (NeurIPS), 2026 · Poster
How Does the Lagrangian Guide Safe Reinforcement Learning through Diffusion Models?
International Conference on Machine Learning (ICML), 2026 · Poster
Information Shapes Koopman Representation
International Conference on Learning Representations (ICLR), 2026 · Oral
Physics-guided multistage neural network: A physically guided network for step initial values and dispersive shock wave phenomena
Physical Review E, 2024
MSTCN-VAE: An unsupervised learning method for micro gesture recognition based on skeleton modality
International Joint Conference on Artificial Intelligence (IJCAI), 2023