<ul><li>Algorithm selection is crucial for identifying the best algorithm for a given problem in continuous black-box optimization.</li><li>Using a set of features to train machine learning meta-models for algorithm selection has been effective, but evaluation approaches need improvement.</li><li>Issues with the 'leave-instance-out' evaluation technique and the impact of objective function scale on performance metrics are highlighted.</li><li>This study stresses the importance of careful evaluation in algorithm selection to avoid misleading results and noise in the field.</li></ul>

The Pitfalls of Benchmarking in Algorithm Selection: What We Are Getting Wrong

Discover more