Orthogonally Decoupled Variational Fourier Features
This work addresses the challenge of efficient Gaussian process modeling for large datasets, but it is incremental as it builds on existing spectral and sparse methods.
The paper tackled the problem of scaling Gaussian processes to big data by combining spectral methods with sparse inducing points using an orthogonally decoupled variational basis, achieving competitive performance with state-of-the-art methods on synthetic and real-world data.
Sparse inducing points have long been a standard method to fit Gaussian processes to big data. In the last few years, spectral methods that exploit approximations of the covariance kernel have shown to be competitive. In this work we exploit a recently introduced orthogonally decoupled variational basis to combine spectral methods and sparse inducing points methods. We show that the method is competitive with the state-of-the-art on synthetic and on real-world data.