English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Despite all the benefits of automated hyperparameter optimization (HPO), mostmodern HPO algorithms are black-boxes themselves. This makes it difficult tounderstand the decision process which leads to the selected configuration,reduces trust in HPO, and thus hinders its broad adoption. Here, we study thecombination of HPO with interpretable machine learning (IML) methods such aspartial dependence plots. These techniques are more and more used to explainthe marginal effect of hyperparameters on the black-box cost function or toquantify the importance of hyperparameters. However, if such methods arenaively applied to the experimental data of the HPO process in a post-hocmanner, the underlying sampling bias of the optimizer can distortinterpretations. We propose a modified HPO method which efficiently balancesthe search for the global optimum w.r.t. predictive performance \emph{and} thereliable estimation of IML explanations of an underlying black-box function bycoupling Bayesian optimization and Bayesian Algorithm Execution. On benchmarkcases of both synthetic objectives and HPO of a neural network, we demonstratethat our method returns more reliable explanations of the underlying black-boxwithout a loss of optimization performance.

Improving Accuracy of Interpretability Measures in Hyperparameter Optimization via Bayesian Algorithm Execution