Learning spectrograms with convolutional spectral kernels

Loading...
Thumbnail Image

Access rights

openAccess

URL

Journal Title

Journal ISSN

Volume Title

A4 Artikkeli konferenssijulkaisussa

Date

2020

Department

Department of Computer Science

Major/Subject

Mcode

Degree programme

Language

en

Pages

3826-3836

Series

The 23rd International Conference on Artificial Intelligence and Statistics, Proceedings of Machine Learning Research, Volume 108

Abstract

We introduce the convolutional spectral kernel (CSK), a novel family of non-stationary, nonparametric covariance kernels for Gaussian process (GP) models, derived from the convolution between two imaginary radial basis functions. We present a principled framework to interpret CSK, as well as other deep probabilistic models, using approximated Fourier transform, yielding a concise representation of input-frequency spectrogram. Observing through the lens of the spectrogram, we provide insight on the interpretability of deep models. We then infer the functional hyperparameters using scalable variational and MCMC methods. On small- and medium-sized spatiotemporal datasets, we demonstrate improved generalization of GP models when equipped with CSK, and their capability to extract non-stationary periodic patterns.

Description

Keywords

Other note

Citation

Shen, Z, Heinonen, M & Kaski, S 2020, Learning spectrograms with convolutional spectral kernels . in S Chiappa & R Calandra (eds), The 23rd International Conference on Artificial Intelligence and Statistics . Proceedings of Machine Learning Research, vol. 108, JMLR, pp. 3826-3836, International Conference on Artificial Intelligence and Statistics, Palermo, Italy, 03/06/2020 . < http://proceedings.mlr.press/v108/shen20a.html >