An Upper Limit of Decaying Rate with Respect to Frequency in Deep Neural Network

Abstract

Deep neural network (DNN) usually learns the target function from low to high frequency, which is called frequency principle or spectral bias. This frequency principle sheds light on a high-frequency curse of DNNs -- difficult to learn high-frequency information. Inspired by the frequency principle, a series of works are devoted to develop algorithms for overcoming the high-frequency curse. A natural question arises: what is the upper limit of the decaying rate w.r.t. frequency when one trains a DNN? In this work, our theory, confirmed by numerical experiments, suggests that there is a critical decaying rate w.r.t. frequency in DNN training. Below the upper limit of the decaying rate, the DNN interpolates the training data by a function with a certain regularity. However, above the upper limit, the DNN interpolates the training data by a trivial function, i.e., a function is only non-zero at training data points. Our results indicate a better way to overcome the high-frequency curse is to design a proper pre-condition approach to shift high-frequency information to low-frequency one, which coincides with several previous developed algorithms for fast learning high-frequency information. More importantly, this work rigorously proves that the high-frequency curse is an intrinsic difficulty of DNNs.

arXiv admin note: substantial text overlap with arXiv:2012.03238

References in corpus (10)

• Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains
• Multi-scale Deep Neural Network (MscaleDNN) for Solving Poisson-Boltzmann Equation in Complex Domains
• Neural Additive Models: Interpretable Machine Learning with Neural Nets
• Explicitizing an Implicit Bias of the Frequency Principle in Two-layer Neural Networks
• Accelerating Physics-Informed Neural Network Training with Prior Dictionaries
• Multi-scale Deep Neural Networks for Solving High Dimensional PDEs
• Phase diagram for two-layer ReLU neural networks at infinite-width limit
• Learned Initializations for Optimizing Coordinate-Based Neural Representations
• Reproducing Activation Function for Deep Learning
• Band-limited Soft Actor Critic Model