Skip to yearly menu bar Skip to main content


PriorBand: Practical Hyperparameter Optimization in the Age of Deep Learning

Neeratyoy Mallik · Edward Bergman · Carl Hvarfner · Danny Stoll · Maciej Janowski · Marius Lindauer · Luigi Nardi · Frank Hutter

Great Hall & Hall B1+B2 (level 1) #1302
[ ]
[ Paper [ Poster [ OpenReview
Thu 14 Dec 3 p.m. PST — 5 p.m. PST


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.

Chat is not available.