Self-adaptive deep neural network: Numerical approximation to functions and PDEs
arXiv:2109.02839 · doi:10.1016/j.jcp.2022.111021
Abstract
Designing an optimal deep neural network for a given task is important and challenging in many machine learning applications. To address this issue, we introduce a self-adaptive algorithm: the adaptive network enhancement (ANE) method, written as loops of the form train, estimate and enhance. Starting with a small two-layer neural network (NN), the step train is to solve the optimization problem at the current NN; the step estimate is to compute a posteriori estimator/indicators using the solution at the current NN; the step enhance is to add new neurons to the current NN. Novel network enhancement strategies based on the computed estimator/indicators are developed in this paper to determine how many new neurons and when a new layer should be added to the current NN. The ANE method provides a natural process for obtaining a good initialization in training the current NN; in addition, we introduce an advanced procedure on how to initialize newly added neurons for a better approximation. We demonstrate that the ANE method can automatically design a nearly minimal NN for learning functions exhibiting sharp transitional layers as well as discontinuous solutions of hyperbolic partial differential equations.
Published in Journal of Computational Physics
References in corpus (5)
- Natural Language Processing (almost) from Scratch
- Least-Squares ReLU Neural Network (LSNN) Method For Linear Advection-Reaction Equation
- Least-Squares ReLU Neural Network (LSNN) Method For Scalar Nonlinear Hyperbolic Conservation Law
- Adaptive Two-Layer ReLU Neural Network: II. Ritz Approximation to Elliptic PDEs
- Adaptive Two-Layer ReLU Neural Network: I. Best Least-squares Approximation