Scalable Gaussian Processes for Characterizing Multidimensional Change Surfaces
This work addresses the challenge of analyzing complex, large-scale data for applications like disease monitoring, though it appears incremental as it builds on existing Gaussian process and kernel methods.
The authors tackled the problem of identifying smooth multidimensional changepoints in large datasets by developing a scalable Gaussian process model, which they demonstrated on a large spatio-temporal disease dataset to uncover previously unknown heterogeneous changes.
We present a scalable Gaussian process model for identifying and characterizing smooth multidimensional changepoints, and automatically learning changes in expressive covariance structure. We use Random Kitchen Sink features to flexibly define a change surface in combination with expressive spectral mixture kernels to capture the complex statistical structure. Finally, through the use of novel methods for additive non-separable kernels, we can scale the model to large datasets. We demonstrate the model on numerical and real world data, including a large spatio-temporal disease dataset where we identify previously unknown heterogeneous changes in space and time.