Towards Safe Multi-Task Bayesian Optimization
This work addresses safety in online optimization for systems where hyperparameters are uncertain, which is an incremental improvement over existing methods that assume known hyperparameters.
The paper tackles the problem of safe Bayesian optimization when hyperparameters are unknown by extending robust Gaussian process error bounds to a multi-task setting, using Markov chain Monte Carlo to compute confidence regions from hyperparameter posteriors, and simulation results show it significantly accelerates optimization for expensive functions compared to state-of-the-art methods, depending on model fidelity.
Bayesian optimization has emerged as a highly effective tool for the safe online optimization of systems, due to its high sample efficiency and noise robustness. To further enhance its efficiency, reduced physical models of the system can be incorporated into the optimization process, accelerating it. These models are able to offer an approximation of the actual system, and evaluating them is significantly cheaper. The similarity between the model and reality is represented by additional hyperparameters, which are learned within the optimization process. Safety is a crucial criterion for online optimization methods such as Bayesian optimization, which has been addressed by recent works that provide safety guarantees under the assumption of known hyperparameters. In practice, however, this does not apply. Therefore, we extend the robust Gaussian process uniform error bounds to meet the multi-task setting, which involves the calculation of a confidence region from the hyperparameter posterior distribution utilizing Markov chain Monte Carlo methods. Subsequently, the robust safety bounds are employed to facilitate the safe optimization of the system, while incorporating measurements of the models. Simulation results indicate that the optimization can be significantly accelerated for expensive to evaluate functions in comparison to other state-of-the-art safe Bayesian optimization methods, contingent on the fidelity of the models.