Log In Sign Up

Infinite-Horizon Gaussian Processes

by   Arno Solin, et al.

Gaussian processes provide a flexible framework for forecasting, removing noise, and interpreting long temporal datasets. State space modelling (Kalman filtering) enables these non-parametric models to be deployed on long datasets by reducing the complexity to linear in the number of data points. The complexity is still cubic in the state dimension m which is an impediment to practical application. In certain special cases (Gaussian likelihood, regular spacing) the GP posterior will reach a steady posterior state when the data are very long. We leverage this and formulate an inference scheme for GPs with general likelihoods, where inference is based on single-sweep EP (assumed density filtering). The infinite-horizon model tackles the cubic cost in the state dimensionality and reduces the cost in the state dimension m to O(m^2) per data point. The model is extended to online-learning of hyperparameters. We show examples for large finite-length modelling problems, and present how the method runs in real-time on a smartphone on a continuous data stream updated at 100 Hz.


Combining Pseudo-Point and State Space Approximations for Sum-Separable Gaussian Processes

Gaussian processes (GPs) are important probabilistic tools for inference...

State Space Gaussian Processes with Non-Gaussian Likelihood

We provide a comprehensive overview and tooling for GP modeling with non...

Kalman Filtering with Gaussian Processes Measurement Noise

Real-world measurement noise in applications like robotics is often corr...

Streaming Variational Monte Carlo

Nonlinear state-space models are powerful tools to describe dynamical st...

A Greedy approximation scheme for Sparse Gaussian process regression

In their standard form Gaussian processes (GPs) provide a powerful non-p...

Numerical Gaussian process Kalman filtering

Numerical Gaussian processes have recently been developed to handle spat...

High-Dimensional Gaussian Process Inference with Derivatives

Although it is widely known that Gaussian processes can be conditioned o...

Code Repositories


Infinite-horizon Gaussian processes

view repo