SigGPDE: Scaling Sparse Gaussian Processes on Sequential Data

Maud Lemercier 1 Cristopher Salvi 2 Thomas Cass 3 Edwin V. Bonilla 4 Theodoros Damoulas 1 Terry Lyons 2

             Abstract                 vations N , with nave approaches having a time complexity
                                  O(N 3 ). From a wide range of approximate techniques
   Making predictions and quantifying their uncer-        to scale inference in GP models to large datasets, “sparse”
   tainty when the input data ...