Fantasizing with Dual GPs in Bayesian Optimization and Active Learning
This work addresses a bottleneck for researchers and practitioners in sequential modeling, offering an incremental improvement over existing sparse models.
The paper tackles the computational inefficiency of Gaussian processes in batch acquisition for Bayesian optimization and active learning by introducing a sparse Dual GP parameterization, achieving linear scaling with batch size and one-step updates for non-Gaussian likelihoods.
Gaussian processes (GPs) are the main surrogate functions used for sequential modelling such as Bayesian Optimization and Active Learning. Their drawbacks are poor scaling with data and the need to run an optimization loop when using a non-Gaussian likelihood. In this paper, we focus on `fantasizing' batch acquisition functions that need the ability to condition on new fantasized data computationally efficiently. By using a sparse Dual GP parameterization, we gain linear scaling with batch size as well as one-step updates for non-Gaussian likelihoods, thus extending sparse models to greedy batch fantasizing acquisition functions.