Fourier Features Let Networks Learn High Frequency Functions in Low Dimensional Domains
arXiv:2006.10739
Abstract
We show that passing input points through a simple Fourier feature mapping enables a multilayer perceptron (MLP) to learn high-frequency functions in low-dimensional problem domains. These results shed light on recent advances in computer vision and graphics that achieve state-of-the-art results by using MLPs to represent complex 3D objects and scenes. Using tools from the neural tangent kernel (NTK) literature, we show that a standard MLP fails to learn high frequencies both in theory and in practice. To overcome this spectral bias, we use a Fourier feature mapping to transform the effective NTK into a stationary kernel with a tunable bandwidth. We suggest an approach for selecting problem-specific Fourier features that greatly improves the performance of MLPs for low-dimensional regression tasks relevant to the computer vision and graphics communities.
Project page: https://people.eecs.berkeley.edu/~bmild/fourfeat/
References in corpus (6)
- Fine-Grained Analysis of Optimization and Generalization for Overparameterized Two-Layer Neural Networks
- Learning to Infer Implicit Surfaces without 3D Supervision
- Neural Tangents: Fast and Easy Infinite Neural Networks in Python
- Time2Vec: Learning a Vector Representation of Time
- Self-attention with Functional Time Representation Learning
- Compressive sensing with un-trained neural networks: Gradient descent finds the smoothest approximation
Cited by in corpus (10)
- When and why PINNs fail to train: A neural tangent kernel perspective
- NVIDIA SimNet^{TM}: an AI-accelerated multi-physics simulation framework
- FreeStyleGAN: Free-view Editable Portrait Rendering with the Camera Manifold
- Unsupervised part representation by Flow Capsules
- Reproducing Activation Function for Deep Learning
- Transformers for One-Shot Visual Imitation
- Physics Informed Deep Learning for Flow and Transport in Porous Media
- Moser Flow: Divergence-based Generative Modeling on Manifolds
- Convolutional Neural Opacity Radiance Fields
- Realistic Image Synthesis with Configurable 3D Scene Layouts