Comparable Machine Learning Efficiency : Balanced Metrics for Natural Language Processing
- As machine learning becomes increasingly pervasive, its resource demands and financial implications escalate, necessitating energy and cost optimisations to meet stakeholder demands. Quality metrics for predictive machine learning models are abundant, but efficiency metrics remain rare. We propose a framework for efficiency metrics, that enables the comparison of distinct efficiency types. A quality-focused efficiency metric is introduced that considers resource consumption, computational effort, and runtime in addition to prediction quality. The metric has been successfully tested for usability, plausibility, and compensation for dataset size and host performance. This framework enables informed decisions to be made about the use and design of machine learning in an environmentally responsible and cost-effective manner.
| Author: | Daniel SchönleORCiD, Christoph ReichORCiDGND, Djaffar Ould-Abdeslam |
|---|---|
| URL: | http://www.thinkmind.org/index.php?view=article&articleid=green_2023_1_30_80014 |
| ISBN: | 978-1-68558-097-1 |
| Parent Title (English): | GREEN 2023 : The Eighth International Conference on Green Communications, Computing and Technologies, September 25 - 29, 2023, Porto, Portugal |
| Publisher: | IARIA |
| Document Type: | Conference Proceeding |
| Language: | English |
| Year of Completion: | 2023 |
| Release Date: | 2024/02/15 |
| Tag: | Automl; Machine learning; Metric; NLP |
| First Page: | 15 |
| Last Page: | 24 |
| Licence (German): |  Urheberrechtlich geschützt |