Hyperparameters of Deep Learning (DL) pipelines are crucial for their downstream performance. While a large number of methods for Hyperparameter Optimization (HPO) have been developed, their incurred costs are often untenable for modern DL. Consequently, manual experimentation is still the most prevalent approach to optimize hyperparameters, relying on the researcher's intuition, domain knowledge, and cheap preliminary explorations. To resolve this misalignment between HPO algorithms and DL researchers, we propose PriorBand, an HPO algorithm tailored to DL, able to utilize both expert beliefs and cheap proxy tasks. Empirically, we demonstrate PriorBand's efficiency across a range of DL benchmarks and show its gains under informative expert input and robustness against poor expert beliefs.

We present mlr3mbo, a comprehensive and modular toolbox for Bayesian optimization in R. mlr3mbo supports single- and multi-objective optimization, multi-point proposals, batch and asynchronous parallelization, input and output transformations, and robust error handling. While it can be used for many standard Bayesian optimization variants in applied settings, researchers can also construct custom BO algorithms from its flexible building blocks. In addition to an introduction to the software, its design principles, and its building blocks, the paper presents two extensive empirical evaluations of the software on the surrogate-based benchmark suite YAHPO Gym. To identify robust default configurations for both numeric and mixed-hierarchical optimization regimes, and to gain further insights into the respective impacts of individual settings, we run a coordinate descent search over the mlr3mbo configuration space and analyze its results. Furthermore, we demonstrate that mlr3mbo achieves state-of-the-art performance by benchmarking it against a wide range of optimizers, including HEBO, SMAC3, Ax, and Optuna.

In recent years, research on Automated Machine Learning (AutoML) [1] has made great strides in the data-driven design of neural network architectures [2, 3] and training hyperparameters [4].

Neural Information Processing Systems http://nips.cc/

The model combines aprobabilistic encoder withamulti-task model suchthatitcan generate inexpensive and realistic tasks of the class of problems of interest.

Weempirically assess theimpact oftheseregularization cocktailsforMLPs ina large-scale empirical study comprising 40 tabular datasets and demonstrate that (i) well-regularized plain MLPs significantly outperform recent state-of-the-art specialized neural network architectures, and (ii) they even outperform strong traditionalMLmethods,suchasXGBoost.

Computational modeling and machine learning (ML) have led to an acceleration of scientific innovation in diverse areas, ranging from drug design to robotics to material science.

Toaddress theabovechallenges, wepropose anovelhyperparameter mutation (HPM) scheduling algorithm in this study, which adopts a population based training framework to explicitly learn a trade-off (i.e., a mutation schedule) between using the hypergradient-guided local search and the mutation-driven global search.

Neural Architecture Search (NAS) wasfirst proposed toachievestate-of-the-art performance through thediscoveryofnewarchitecture patterns, without human intervention.