Finding the SWEET Spot: Analysis and Improvement of Adaptive Inference in Low Resource Settings

Daniel Rotem, Michael Hassid, Jonathan Mamou, Roy Schwartz

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Adaptive inference is a simple method for reducing inference costs. The method works by maintaining multiple classifiers of different capacities, and allocating resources to each test instance according to its difficulty. In this work, we compare the two main approaches for adaptive inference, Early-Exit and Multi-Model, when training data is limited. First, we observe that for models with the same architecture and size, individual Multi-Model classifiers outperform their Early-Exit counterparts by an average of 2.3%. We show that this gap is caused by Early-Exit classifiers sharing model parameters during training, resulting in conflicting gradient updates of model weights. We find that despite this gap, Early-Exit still provides a better speed-accuracy trade-off due to the overhead of the Multi-Model approach. To address these issues, we propose SWEET, an Early-Exit fine-tuning method that assigns each classifier its own set of unique model weights, not updated by other classifiers. We compare SWEET's speed-accuracy curve to standard Early-Exit and Multi-Model baselines and find that it outperforms both methods at fast speeds while maintaining comparable scores to Early-Exit at slow speeds. Moreover, SWEET individual classifiers outperform Early-Exit ones by 1.1% on average. SWEET enjoys the benefits of both methods, paving the way for further reduction of inference costs in NLP. We publicly release our code.

Original languageEnglish
Title of host publicationLong Papers
PublisherAssociation for Computational Linguistics (ACL)
Pages14836-14851
Number of pages16
ISBN (Electronic)9781959429722
StatePublished - 2023
Event61st Annual Meeting of the Association for Computational Linguistics, ACL 2023 - Toronto, Canada
Duration: 9 Jul 202314 Jul 2023

Publication series

NameProceedings of the Annual Meeting of the Association for Computational Linguistics
Volume1

Conference

Conference61st Annual Meeting of the Association for Computational Linguistics, ACL 2023
Country/TerritoryCanada
CityToronto
Period9/07/2314/07/23

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Linguistics and Language
  • Language and Linguistics

Fingerprint

Dive into the research topics of 'Finding the SWEET Spot: Analysis and Improvement of Adaptive Inference in Low Resource Settings'. Together they form a unique fingerprint.

Cite this