Publications

Working Papers

• Learning coarse-grained dynamics with causal and conservative spatiotemporal neural operators.
  Junfeng Chen.
  

• Spectrally constrained discovery resolves the stability-expressivity paradox in learning stiff multiscale physics.
  Qinghe Wang, Junfeng Chen, Kailiang Wu.
  submitted 2026.
  

• Continuous tokenizer for scientific AI: Unifying discrete data and continuous physical dynamics.
  Junfeng Chen, Qinghe Wang, Kailiang Wu.
  submitted 2026.
  

Peer Reviewed

• DUE: A deep learning framework and library for modeling unknown equations.
  Junfeng Chen, Kailiang Wu, Dongbin Xiu.
  SIAM Review 2025.
  [Paper] [Code]

• Positional knowledge is all you need: Position-induced Transformer (PiT) for operator learning.
  Junfeng Chen, Kailiang Wu.
  International Conference on Machine Learning (ICML) 2024.
  [Paper] [Poster] [Code]

• Deep-OSG: Deep learning of operators in semigroup.
  Junfeng Chen, Kailiang Wu.
  Journal of Computational Physics 2023.
  [Paper]

• A twin-decoder structure for incompressible laminar flow reconstruction with uncertainty estimation around 2D obstacles.
  Junfeng Chen, Jonathan Viquerat, Frédéric Heymes, Elie Hachem.
  Neural Computing and Applications 2022.
  [Paper] [Code]

• Graph neural networks for laminar flow prediction around random two-dimensional shapes.
  Junfeng Chen, Elie Hachem, Jonathan Viquerat.
  Physics of Fluids 2021.
  [Paper] [Code]

Others

• U-net architectures for fast prediction of incompressible laminar flows.
  Junfeng Chen, Jonathan Viquerat, Elie Hachem.
  arxiv 2019
  [paper]