Graph Convolutional Gaussian Processes
This addresses the need for flexible models in machine learning for data on graphs, such as in image processing or mesh analysis, but it appears incremental as it combines existing concepts like graph convolutions and Gaussian processes.
The authors tackled the problem of learning translation-invariant relationships on non-Euclidean domains by proposing graph convolutional Gaussian processes, a Bayesian nonparametric method that can handle high-dimensional inputs on general graphs, and demonstrated favorable comparisons to existing methods in applications like images and triangular meshes.
We propose a novel Bayesian nonparametric method to learn translation-invariant relationships on non-Euclidean domains. The resulting graph convolutional Gaussian processes can be applied to problems in machine learning for which the input observations are functions with domains on general graphs. The structure of these models allows for high dimensional inputs while retaining expressibility, as is the case with convolutional neural networks. We present applications of graph convolutional Gaussian processes to images and triangular meshes, demonstrating their versatility and effectiveness, comparing favorably to existing methods, despite being relatively simple models.