DEHB: Evolutionary Hyperband for Scalable, Robust and Efficient Hyperparameter Optimization
This addresses the problem of scalable and efficient HPO for machine learning practitioners, offering a potentially new default method with incremental improvements over existing approaches.
The paper tackled hyperparameter optimization (HPO) by combining Hyperband and Differential Evolution into DEHB, resulting in a method that is up to 1000x faster than random search and demonstrates robust performance across diverse HPO problems and neural architecture search benchmarks.
Modern machine learning algorithms crucially rely on several design decisions to achieve strong performance, making the problem of Hyperparameter Optimization (HPO) more important than ever. Here, we combine the advantages of the popular bandit-based HPO method Hyperband (HB) and the evolutionary search approach of Differential Evolution (DE) to yield a new HPO method which we call DEHB. Comprehensive results on a very broad range of HPO problems, as well as a wide range of tabular benchmarks from neural architecture search, demonstrate that DEHB achieves strong performance far more robustly than all previous HPO methods we are aware of, especially for high-dimensional problems with discrete input dimensions. For example, DEHB is up to 1000x faster than random search. It is also efficient in computational time, conceptually simple and easy to implement, positioning it well to become a new default HPO method.