Safe Exploration for Optimization with Gaussian Processes

We consider sequential decision problems under uncertainty, where we seek to optimize an unknown function from noisy samples. This requires balancing exploration (learning about the objective) and exp