On the eigenvector bias of Fourier feature networks: From regression to solving multi-scale PDEs with physics-informed neural networks.
Sifan WangHanwen WangParis PerdikarisPublished in: CoRR (2020)
Keyphrases
- neural network
- multiscale
- partial differential equations
- regression model
- linear combination
- scale space
- fourier transform
- pattern recognition
- complex networks
- artificial neural networks
- neural nets
- fuzzy logic
- image processing
- regression problems
- computer science
- neural network model
- social networks
- neural network ensemble
- frequency domain
- support vector regression
- fourier domain
- supervised feature selection
- feature vectors
- genetic algorithm
- artificial intelligence
- image segmentation
- image features
- edge detection
- natural images
- mathematical morphology
- wavelet transform
- feed forward
- self organizing maps
- gaussian processes
- activation function
- level set
- multiple scales
- computer vision
- recurrent networks
- fourier spectrum
- image reconstruction