Deep Structured Mixtures of Gaussian Processes

Martin Trapp, Robert Peharz, Franz Pernkopf, Carl E. Rasmussen

    Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review


    Gaussian Processes (GPs) are powerful non-parametric Bayesian regression models that allow exact posterior inference, but exhibit high computational and memory costs. In order to improve scalability of GPs, approximate posterior inference is frequently employed, where a prominent class of approximation techniques is based on local GP experts. However, local-expert techniques proposed so far are either not well-principled, come with limited approximation guarantees, or lead to intractable models. In this paper, we introduce deep structured mixtures of GP experts, a stochastic process model which i) allows exact posterior inference, ii) has attractive computational and memory costs, and iii) when used as GP approximation, captures predictive uncertainties consistently better than previous expert-based approximations. In a variety of experiments, we show that deep structured mixtures have a low approximation error and often perform competitive or outperform prior work.
    Original languageEnglish
    Title of host publicationInternational Conference on Artificial Intelligence and Statistics (AISTATS)
    PublisherProceedings of Machine Learning Research
    Publication statusAccepted/In press - 26 Apr 2020
    Event23rd International Conference on Artificial Intelligence and Statistics - Palermo, Italy
    Duration: 3 Jun 20205 Jun 2020


    Conference23rd International Conference on Artificial Intelligence and Statistics
    Abbreviated titleAISTATS 2020

    Bibliographical note

    AISTATS 2020, in press

    Fingerprint Dive into the research topics of 'Deep Structured Mixtures of Gaussian Processes'. Together they form a unique fingerprint.

    Cite this