Timezone: »

 
Sequential Gaussian Processes for Online Learning of Nonstationary Functions
Michael Minyi Zhang · Bianca Dumitrascu · Sinead Williamson · Barbara Engelhardt

We propose a sequential Monte Carlo algorithm to fit infinite mixtures of GPs that capture non-stationary behavior while allowing for online, distributed inference. Our approach empirically improves performance over state-of-the-art methods for online GP estimation in the presence of non-stationarity in time-series data. To demonstrate the utility of our proposed online Gaussian process mixture-of-experts approach in applied settings, we show that we can successfully implement an optimization algorithm using online Gaussian process bandits.

Author Information

Michael Minyi Zhang (University of Hong Kong)
Bianca Dumitrascu (Cambridge University)
Sinead Williamson (University of Texas at Austin)
Barbara Engelhardt (Princeton University)

More from the Same Authors