Learning Specialized Activation Functions for Physics-informed Neural Networks
arXiv:2308.04073 · doi:10.4208/cicp.OA-2023-0058
Abstract
Physics-informed neural networks (PINNs) are known to suffer from optimization difficulty. In this work, we reveal the connection between the optimization difficulty of PINNs and activation functions. Specifically, we show that PINNs exhibit high sensitivity to activation functions when solving PDEs with distinct properties. Existing works usually choose activation functions by inefficient trial-and-error. To avoid the inefficient manual selection and to alleviate the optimization difficulty of PINNs, we introduce adaptive activation functions to search for the optimal function when solving different problems. We compare different adaptive activation functions and discuss their limitations in the context of PINNs. Furthermore, we propose to tailor the idea of learning combinations of candidate activation functions to the PINNs optimization, which has a higher requirement for the smoothness and diversity on learned functions. This is achieved by removing activation functions which cannot provide higher-order derivatives from the candidate set and incorporating elementary functions with different properties according to our prior knowledge about the PDE at hand. We further enhance the search space with adaptive slopes. The proposed adaptive activation function can be used to solve different PDE systems in an interpretable way. Its effectiveness is demonstrated on a series of benchmarks. Code is available at https://github.com/LeapLabTHU/AdaAFforPINNs.
References in corpus (26)
- Neural Tangent Kernel: Convergence and Generalization in Neural Networks
- Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains
- Adaptive activation functions accelerate convergence in deep and physics-informed neural networks
- Gradient-enhanced physics-informed neural networks for forward and inverse PDE problems
- A comprehensive study of non-adaptive and residual-based adaptive sampling for physics-informed neural networks
- Physics-informed neural networks for inverse problems in nano-optics and metamaterials
- Frequency Principle: Fourier Analysis Sheds Light on Deep Neural Networks
- Quantifying total uncertainty in physics-informed neural networks for solving forward and inverse stochastic problems
- A Physics-Informed Machine Learning Approach for Solving Heat Transfer Equation in Advanced Manufacturing and Engineering Applications
- Deep Learning of Vortex Induced Vibrations
- Finite Basis Physics-Informed Neural Networks (FBPINNs): a scalable domain decomposition approach for solving differential equations
- Learning Activation Functions to Improve Deep Neural Networks
- Collapse of Deep and Narrow Neural Nets
- Dying ReLU and Initialization: Theory and Numerical Examples
- A Physics Informed Neural Network for Time-Dependent Nonlinear and Higher Order Partial Differential Equations
- On the convergence of physics informed neural networks for linear second-order elliptic and parabolic type PDEs
- Locally adaptive activation functions with slope recovery term for deep and physics-informed neural networks
- Meta-learning PINN loss functions
- Self-Adaptive Physics-Informed Neural Networks using a Soft Attention Mechanism
- Learning in Sinusoidal Spaces with Physics-Informed Neural Networks
- Learning Combinations of Activation Functions
- Physics-informed neural networks with hard constraints for inverse design
- A rate of convergence of Physics Informed Neural Networks for the linear second order elliptic PDEs
- A multi-scale DNN algorithm for nonlinear elliptic equations with multiple scales
- MOD-Net: A Machine Learning Approach via Model-Operator-Data Network for Solving PDEs
- Is Physics-Informed Loss Always Suitable for Training Physics-Informed Neural Network?