Variational bridge constructs for approximate Gaussian process regression
This provides a scalable approximation method for Gaussian process regression, which is incremental as it builds on existing variational and SDE techniques.
The paper tackles the computational challenge of Gaussian process regression by introducing a variational bridge method that approximates it via stochastic differential equations and variational inference, achieving approximations indistinguishable from full GP samples.
This paper introduces a method to approximate Gaussian process regression by representing the problem as a stochastic differential equation and using variational inference to approximate solutions. The approximations are compared with full GP regression and generated paths are demonstrated to be indistinguishable from GP samples. We show that the approach extends easily to non-linear dynamics and discuss extensions to which the approach can be easily applied.