Spectral Representations for Convolutional Neural Networks

06/11/2015
by   Oren Rippel, et al.
0

Discrete Fourier transforms provide a significant speedup in the computation of convolutions in deep learning. In this work, we demonstrate that, beyond its advantages for efficient computation, the spectral domain also provides a powerful representation in which to model and train convolutional neural networks (CNNs). We employ spectral representations to introduce a number of innovations to CNN design. First, we propose spectral pooling, which performs dimensionality reduction by truncating the representation in the frequency domain. This approach preserves considerably more information per parameter than other pooling strategies and enables flexibility in the choice of pooling output dimensionality. This representation also enables a new form of stochastic regularization by randomized modification of resolution. We show that these methods achieve competitive results on classification and approximation tasks, without using any dropout or max-pooling. Finally, we demonstrate the effectiveness of complex-coefficient spectral parameterization of convolutional filters. While this leaves the underlying model unchanged, it results in a representation that greatly facilitates optimization. We observe on a variety of popular CNN configurations that this leads to significantly faster convergence during training.

READ FULL TEXT

page 3

page 5

page 6

research
10/07/2018

Hartley Spectral Pooling for Deep Learning

In most convolution neural networks (CNNs), downsampling hidden layers i...
research
12/04/2015

Max-Pooling Dropout for Regularization of Convolutional Neural Networks

Recently, dropout has seen increasing use in deep learning. For deep con...
research
06/10/2016

Deep CNNs along the Time Axis with Intermap Pooling for Robustness to Spectral Variations

Convolutional neural networks (CNNs) with convolutional and pooling oper...
research
12/15/2022

Graphon Pooling for Reducing Dimensionality of Signals and Convolutional Operators on Graphs

In this paper we propose a pooling approach for convolutional informatio...
research
11/26/2022

Transform Once: Efficient Operator Learning in Frequency Domain

Spectral analysis provides one of the most effective paradigms for infor...
research
05/06/2020

Regularized Pooling

In convolutional neural networks (CNNs), pooling operations play importa...

Please sign up or login with your details

Forgot password? Click here to reset