LMEMs for post-hoc analysis of HPO Benchmarking

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Geburek, Anton, Mallik, Neeratyoy, Stoll, Danny, Bouthillier, Xavier, Hutter, Frank
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929449674473472
author Geburek, Anton
Mallik, Neeratyoy
Stoll, Danny
Bouthillier, Xavier
Hutter, Frank
author_facet Geburek, Anton
Mallik, Neeratyoy
Stoll, Danny
Bouthillier, Xavier
Hutter, Frank
contents The importance of tuning hyperparameters in Machine Learning (ML) and Deep Learning (DL) is established through empirical research and applications, evident from the increase in new hyperparameter optimization (HPO) algorithms and benchmarks steadily added by the community. However, current benchmarking practices using averaged performance across many datasets may obscure key differences between HPO methods, especially for pairwise comparisons. In this work, we apply Linear Mixed-Effect Models-based (LMEMs) significance testing for post-hoc analysis of HPO benchmarking runs. LMEMs allow flexible and expressive modeling on the entire experiment data, including information such as benchmark meta-features, offering deeper insights than current analysis practices. We demonstrate this through a case study on the PriorBand paper's experiment data to find insights not reported in the original work.
format Preprint
id arxiv_https___arxiv_org_abs_2408_02533
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle LMEMs for post-hoc analysis of HPO Benchmarking
Geburek, Anton
Mallik, Neeratyoy
Stoll, Danny
Bouthillier, Xavier
Hutter, Frank
Machine Learning
The importance of tuning hyperparameters in Machine Learning (ML) and Deep Learning (DL) is established through empirical research and applications, evident from the increase in new hyperparameter optimization (HPO) algorithms and benchmarks steadily added by the community. However, current benchmarking practices using averaged performance across many datasets may obscure key differences between HPO methods, especially for pairwise comparisons. In this work, we apply Linear Mixed-Effect Models-based (LMEMs) significance testing for post-hoc analysis of HPO benchmarking runs. LMEMs allow flexible and expressive modeling on the entire experiment data, including information such as benchmark meta-features, offering deeper insights than current analysis practices. We demonstrate this through a case study on the PriorBand paper's experiment data to find insights not reported in the original work.
title LMEMs for post-hoc analysis of HPO Benchmarking
topic Machine Learning
url https://arxiv.org/abs/2408.02533