Hint

Example code: https://github.com/cselab/korali/tree/master/examples/learning/surrogates/optimization/

Optimization on Surrogates with Gaussian Processes

This tutorial shows how to use a korali surrogate as a model for optimization. The file train-surrogate.py constructs a Gaussian Process surrogate fronm the given synthetic data and saves it inside the _korali_result_surrogate/ folder. Second, we load and use the trained surrogate as a function to optimize. This is done in run-cmaes.py.