About me

I am Dawen Wu (吴达文), a Research Fellow at CNRS@CREATE, working with Prof. Ludovic Chamoin, Prof. Pierre Senellart and Prof. Stephane Bressan. Previously, I was a Ph.D. candidate in Université Paris-Saclay, CNRS, CentraleSupélec, Laboratory of Signals and Systems supervised by Prof. Abdel Lisser.

I develop machine learning-based solvers for a variety of mathematical problems.

Education

• B.S. in Economics and Management, Civil Aviation University of China.
• M.S. in Computer Science, Peking University.
  Advisor: Prof. Zhong Chen.
  Selected coures: 1) Convex optimization, 2) Machine learning, 3) Reinforcement learning
• Ph.D. in Applied Mathematics, Université Paris-Saclay.
  Advisor: Prof. Abdel Lisser.
  Selected coures: 1) Advanced continuous optimization, 2) Optimization and statistical learning, 3) Stability of nonlinear systems.

Selected papers

1. Neural Triangular Map for Density Estimation and Sampling with Application to Bayesian Inference, Dawen Wu, Ludovic Chamoin, Stéphane Bressan, Journal of Computational Physics.

2. Solving Large-Scale Variational Inequalities with Dynamically Adjusting Initial Condition in Physics-Informed Neural Networks, Dawen Wu, Ludovic Chamoin, Abdel Lisser, Computer Methods in Applied Mechanics and Engineering.

3. Neuro-PINN: A Hybrid Framework for Efficient Nonlinear Projection Equation Solutions, Dawen Wu, Abdel Lisser, International Journal for Numerical Methods in Engineering.

4. Separable Hamiltonian Neural Networks, Zi-Yu Khoo, Dawen Wu*, Jonathan Sze Choong Low, Stéphane Bressan, Physical Review E . (*Corresponding author)

5. Parallel Solution of Nonlinear Projection Equations in a Multitask Learning Framework, Dawen Wu, Abdel Lisser, IEEE Transactions on Neural Networks and Learning Systems.

6. Reformulating Chance-Constrained Optimization as Neural Network Learning, Dawen Wu, Ludovic Chamoin, Abdel Lisser, European Journal of Operational Research. (Major Revision)

7. Approximations of the Inverse Cumulative Distribution Function using Transport Maps and Physics-informed Neural Networks, Dawen Wu, Ludovic Chamoin, Inverse Problems. (Reject \& Resubmit)

Prsentations

• National University of Singapore, Singapore, Feb. 07, 2024
  Title: Transport Map. Slide
• Laboratoire des Signaux et Systèmes (L2S) PhD students day, Gif-sur-Yvette, Sep. 15, 2022
  Title: Optimization-informed neural networks. Slide
• ECSO-CMS-2022 Conference, Venice, July 1, 2022
  Title: A dynamical neural network approach for solving stochastic two-player zero-sum games. Slide
• CSC - POLYTECH workshop, Paris, May 23, 2022
  Title: Using cnn for solving two-player zero-sum games. Slide
• Laboratoire des Signaux et Systèmes (L2S) PhD students day, Gif-sur-Yvette, Sep. 23, 2021
  Title: Using cnn for solving two-player zero-sum games. Poster

Academic Service

Invited Reviewer (Journal):

• IEEE Transactions on Neural Networks and Learning Systems
• European Journal of Control
• Journal of Global Optimization
• Machine Learning
• Engineering Applications of Artificial Intelligence
• IEEE Transactions on Cybernetics
• Neural Networks
• Nonlinear Dynamics

Skills

• Programming: Python, Julia.
• Machine Learning: PyTorch, JAX.