ICML 2016poster7 citations

Extended and Unscented Kitchen Sinks

Edwin Bonilla, Daniel Steinberg, Alistair Reid

Abstract

We propose a scalable multiple-output generalization of unscented and extended Gaussian processes. These algorithms have been designed to handle general likelihood models by linearizing them using a Taylor series or the Unscented Transform in a variational inference framework. We build upon random feature approximations of Gaussian process covariance functions and show that, on small-scale single-task problems, our methods can attain similar performance as the original algorithms while having less computational cost. We also evaluate our methods at a larger scale on MNIST and on a seismic inversion which is inherently a multi-task problem.

BibTeX
@InProceedings{pmlr-v48-bonilla16,
  title = 	 {Extended and Unscented Kitchen Sinks},
  author = 	 {Bonilla, Edwin and Steinberg, Daniel and Reid, Alistair},
  booktitle = 	 {Proceedings of The 33rd International Conference on Machine Learning},
  pages = 	 {1651--1659},
  year = 	 {2016},
  editor = 	 {Balcan, Maria Florina and Weinberger, Kilian Q.},
  volume = 	 {48},
  series = 	 {Proceedings of Machine Learning Research},
  address = 	 {New York, New York, USA},
  month = 	 {20--22 Jun},
  publisher =    {PMLR},
  pdf = 	 {http://proceedings.mlr.press/v48/bonilla16.pdf},
  url = 	 {https://proceedings.mlr.press/v48/bonilla16.html},
  abstract = 	 {We propose a scalable multiple-output generalization of unscented and extended Gaussian processes. These algorithms have been designed to handle general likelihood models by linearizing them using a Taylor series or the Unscented Transform in a variational inference framework. We build upon random feature approximations of Gaussian process covariance functions and show that, on small-scale single-task problems, our methods can attain similar performance as the original algorithms while having less computational cost. We also evaluate our methods at a larger scale on MNIST and on a seismic inversion which is inherently a multi-task problem.}
}