Robust Regression with Twinned Gaussian Processes

@inproceedings{Naish2007RobustRW,
  title={Robust Regression with Twinned Gaussian Processes},
  author={Andrew Naish and Sean B. Holden},
  booktitle={NIPS},
  year={2007}
}
We propose a Gaussian process (GP) framework for robust infe renc in which a GP prior on the mixing weights of a two-component noise model augments the standard process over latent function values. This approac h is a generalization of the mixture likelihood used in traditional robust GP regres sion, and a specialization of the GP mixture models suggested by Tresp [1] and Rasmu s en and Ghahramani [2]. The value of this restriction is in its tractable ex p ctation propagation updates, which… CONTINUE READING

References

Publications referenced by this paper.
Showing 1-10 of 12 references

Multivariate adaptive regression splin es.Annals of Statistics

  • J. H. Friedman
  • 1991
Highly Influential
5 Excerpts

Gaussian process models for robust regression, classificat ion nd reinforcement learning

  • Malte Kuss
  • PhD thesis, Technische Universität Darmstadt,
  • 2006
2 Excerpts

Similar Papers

Loading similar papers…