# Deep convolutional Gaussian processes

Source: [Deep convolutional Gaussian processes](<https://www.sota2.com/research/paper/68fa64e11e209131cea0704f>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2025-10-25T21:13:58.604Z

Authors: Kenneth Blomqvist, Samuel Kaski, Markus Heinonen

Published: 2018-10-06T20:58:05.000Z

## 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.

## Recorded benchmark results

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

- [Classification on MNIST (Error Rate, ECE)](<https://www.sota2.com/research/sota/classification-on-mnist-error-rate-ece.md>): Classification; MNIST; Error Rate: 0.56.

- [Classification on CIFAR10 (Error rate, ECE)](<https://www.sota2.com/research/sota/classification-on-cifar10-error-rate-ece.md>): Classification; CIFAR10; Error Rate: 24.1.

## Original sources and implementation

[Original paper](<https://arxiv.org/abs/1810.03052>)

[PDF](<https://arxiv.org/pdf/1810.03052.pdf>)