Yunfeng Liao, an undergraduate researcher, works for Harbin Institute of Technology (Shenzhen). He is now focusing on AI4PDE and further for gaming engines/graphics, under guidance of Assistant Professor Xiucheng Li.

Education

Undergraduate - Computer Science in Harbin Institute of Technology(Shenzhen)

  • GPA: 3.85/4.0 (90.906/100.000).
  • TOFEL: 104/120(R26/L28/S25/W25); CET-6: 586/710; CET-4: 654/710

Publication

(ICML'25) Curvature-aware Graph Attention for PDEs on Manifolds

  • Authors: Yunfeng Liao, Jiawen Guan, Xiucheng Li
  • Status: Accepted

(NeurlPS'25) Boundary-Value PDEs Meet Higher-Order Differential Topology-aware GNNs

  • Authors: Yunfeng Liao, Yangxin Wu, Xiucheng Li
  • Status: Spotlight
  • Region: AI4PDE, Differential Geometry, Electromagnetism, Graph Neural Networks.

Research Interest

  • AI for PDE. He is exploring neural approaches to naturally integrate science (pure math & mathematical physics) and concrete applications (PDE solving). Recently, he devotes to exploring neural methods to solve PDEs in curved spaces or embedded submanifolds.

  • AI for gaming engines/graphics (on physics simulation rendering). Physics simulations (fabrics, particles and fluid simulation, ray-tracing etc) can be time-consuming in rendering. Neural methods (AI4PDE) enjoy high efficiency and tolerable inaccuracy, outperforming numerical solvers in time-sensitive scenarios, such as gaming engines (real-time rendering).

  • Machine learning / Deep learning theory.

  • Genshin Impact. Kaedehara Kazuha!

Skills

  • Pure Math. Advanced Calculus, Real/Complex/Functional/Convex Analysis, Advanced/Abstract/Lie Algebra, Group Representation, Differential Geometry, Basic Topology, Ordinary/Partial Differential Equations.
  • Physics. Analytical/Quatumatic mechanism, Fluid Dynamics, Electrodynamics.
  • Applied Math. Random process, Stochastic Differential Equations, Numerical methods, Discrete Differential Geometry, Computational Geometry, Classical/Bayesian Statistics.
  • Programming: PyTorch, C/C++, JAVA.
  • AI: Deep Learning (Graph Neural Networks, Time Series Analysis), Deep Learning Theory, Reinforce Learning.
  • Learning: Algebraic topology, Geometry Analysis, Ricci Flow, Gauge theory, Graphics, Optimal Transport, Diffusion, OpenGL.