Safe Explorative Bayesian Optimization - Towards Personalized Treatments in Plasma Medicine
Kimberly J. Chan, Joel A. Paulson, Ali Mesbah · 2023
This paper considers the problem of Bayesian optimization (BO) for systems with safety-critical constraints. Recent work has shown that a theoretically consistent way to account for constraints in BO is to relax the constraint functions such that the feasible region has a high probability of containing the global solution. However, by construction, these approaches are unable to ensure safe/feasible operation at every query, which is unacceptable in safety-critical applications. Alternatively, safe BO methods force the query points to remain in the interior of a partially-revealed safety region, which may result in unacceptable (and unquantified) performance losses. This paper presents a new safe BO method that avoids these performance losses by systematically incorporating potential performance gains from enlargement of the safety region. The proposed method avoids getting stuck at suboptimal points based on a potentially small initial safety region due to limited initial exploration of the safety boundary. The performance of the proposed method is demonstrated for safe control of a cold atmospheric plasma jet towards personalized plasma medicine.