Incorporating Continuous Dependence Qualifies Physics-Informed Neural Networks for Operator Learning
This addresses generalization issues in PINNs for solving PDEs, offering a simple improvement for operator learning applications, though it is incremental as it builds on existing PINN methods.
The paper tackled the poor generalization of physics-informed neural networks (PINNs) in operator learning by incorporating continuous dependence of PDE solutions, achieving 1-3 orders of magnitude lower test MSE than DeepONet and FNO with limited labeled data.
Physics-informed neural networks (PINNs) have been proven as a promising way for solving various partial differential equations, especially high-dimensional ones and those with irregular boundaries. However, their capabilities in real applications are highly restricted by their poor generalization performance. Inspired by the rigorous mathematical statements on the well-posedness of PDEs, we develop a novel extension of PINNs by incorporating the additional information on the continuous dependence of PDE solutions with respect to parameters and initial/boundary values (abbreviated as cd-PINN). Extensive numerical experiments demonstrate that, with limited labeled data, cd-PINN achieves 1-3 orders of magnitude lower in test MSE than DeepONet and FNO. Therefore, incorporating the continuous dependence of PDE solutions provides a simple way for qualifying PINNs for operator learning.