The XJTU AI for Scientific Computing Lab develops artificial intelligence methods, randomized neural networks, operator learning models, and structure-preserving numerical algorithms for scientific computing.
- Randomized neural networks for differential equations
- Local randomized neural networks and domain decomposition methods
- Adaptive and growing randomized neural networks
- Operator learning for parametric PDEs
- Structure-preserving AI methods for physical systems
- Applications in neutron transport, electromagnetics, fluid dynamics, electrochemical systems, and multiphysics modeling
- rann-core: shared utilities for randomized neural network methods
- research-code-index: index of paper codes and reproducible experiments
- rann-paper-code-template: template for paper code repositories
- operator-learning: operator learning models and benchmarks
- teaching-benchmarks: tutorial codes for students
- RaNN-Petrov-Galerkin methods
- Local randomized neural networks with DG formulations
- Local randomized neural networks with FDM formulations
- Adaptive and growing randomized neural networks
- Randomized neural operator learning
- AI-assisted scientific computing education
Fei Wang, School of Mathematics and Statistics, Xi'an Jiaotong University.