Deep convolutional Gaussian process

Kenneth Blomqvist, Samuel Kaski, Markus Heinonen

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

Abstract

We propose deep convolutional Gaussian processes, a deep Gaussian process architecture with convolutional structure. The model is a principled Bayesian framework for detecting hierarchical combinations of local features for image classification. We demonstrate greatly improved image classification performance compared to current Gaussian process approaches on the MNIST and CIFAR-10 datasets. In particular, we improve CIFAR-10 accuracy by over 10 percentage points.
Original languageEnglish
Title of host publicationProceedings of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases
Publication statusAccepted/In press - 2019
MoE publication typeA4 Article in a conference publication
EventEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases - Wurzburg, Germany
Duration: 16 Sep 201920 Sep 2019
Conference number: 16

Conference

ConferenceEuropean Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases
Abbreviated titleECML PKDD
CountryGermany
CityWurzburg
Period16/09/201920/09/2019

Keywords

  • gaussian process
  • Deep Learning
  • Bayesian classification

Fingerprint Dive into the research topics of 'Deep convolutional Gaussian process'. Together they form a unique fingerprint.

  • Cite this

    Blomqvist, K., Kaski, S., & Heinonen, M. (Accepted/In press). Deep convolutional Gaussian process. In Proceedings of the European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases