What is it about?
CMA-ES is a very popular optimisation algorithm with many hyper-parameters and extensions. In this work we explore which of these variants are biased to a specific part of the search space and how this structural bias influences the performance of these algorithms.
Featured Image
Photo by Paul Green on Unsplash
Why is it important?
It is important to understand why and how algorithms can be biased. The bias is most of the time an unwanted effect, while sometimes it can be beneficial if the algorithm is biased towards the global optimum.
Read the Original
This page is a summary of: Using structural bias to analyse the behaviour of modular CMA-ES, July 2022, ACM (Association for Computing Machinery),
DOI: 10.1145/3520304.3534035.
You can read the full text:
Resources
Contributors
The following have contributed to this page