% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@INPROCEEDINGS{Aach:1022355,
author = {Aach, Marcel and Inanc, Eray and Sarma, Rakesh and Riedel,
Morris and Lintermann, Andreas},
title = {{O}ptimal {R}esource {A}llocation for {E}arly
{S}topping-based {N}eural {A}rchitecture {S}earch {M}ethods},
volume = {228},
publisher = {PMLR},
reportid = {FZJ-2024-01461},
series = {Proceedings of Machine Learning Research},
pages = {12/1--17},
year = {2023},
abstract = {The field of NAS has been significantly benefiting from the
increased availability of parallel compute resources, as
optimization algorithms typically require sampling and
evaluating hundreds of model configurations. Consequently,
to make use of these resources, the most commonly used early
stopping-based NAS methods are suitable for running multiple
trials in parallel. At the same time, also the training time
of single model configurations can be reduced, e.g., by
employing data-parallel training using multiple GPUs. This
paper investigates the optimal allocation of a fixed amount
of parallel workers for conducting NAS. In practice, users
have to decide if the computational resources are primarily
used to assign more workers to the training of individual
trials or to increase the number of trials executed in
parallel. The first option accelerates the speed of the
individual trials (exploitation) but reduces the parallelism
of the NAS loop, whereas for the second option, the runtime
of the trials is longer but a larger number of
simultaneously processed trials in the NAS loop is achieved
(exploration). Our study encompasses both large- and
small-scale scenarios, including tuning models in parallel
on a single GPU, with data-parallel training on up to 16
GPUs, and measuring the scalability of NAS on up to 64 GPUs.
Our empirical results using the HyperBand, Asynchronous
Successive Halving, and Bayesian Optimization HyperBand
methods offer valuable insights for users seeking to run NAS
on both small and large computational budgets. By selecting
the appropriate number of parallel evaluations, the NAS
process can be accelerated by factors of ${\approx}$2–5
while preserving the test set accuracy compared to
non-optimal resource allocations.}},
month = {Nov},
date = {2023-11-12},
organization = {Second International Conference on
Automated Machine Learning, Potsdam
(Germany), 12 Nov 2023 - 15 Nov 2023},
cin = {JSC},
cid = {I:(DE-Juel1)JSC-20090406},
pnm = {5111 - Domain-Specific Simulation $\&$ Data Life Cycle Labs
(SDLs) and Research Groups (POF4-511) / RAISE - Research on
AI- and Simulation-Based Engineering at Exascale (951733)},
pid = {G:(DE-HGF)POF4-5111 / G:(EU-Grant)951733},
typ = {PUB:(DE-HGF)8 / PUB:(DE-HGF)7},
doi = {10.34734/FZJ-2024-01461},
url = {https://juser.fz-juelich.de/record/1022355},
}