Energy-Efficient Gaussian Processes Using Low-Precision Arithmetic

Nicolas Alder, Ralf Herbrich
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:955-975, 2024.

Abstract

The widespread use of artificial intelligence requires finding energy-efficient paradigms for the field. We propose to reduce the energy consumption of Gaussian process regression using low-precision floating-point representations. We explore how low-precision representations impact the results of Gaussian process regression and how data set properties, implementation approach, model performance, and energy consumption interact. Our findings show that a well-conditioned kernel matrix allows reducing the energy consumption by up to 89.01% for 98.08% of arithmetic operations with little to no impact on model performance. Our findings are relevant whenever one needs to invert a symmetric full-rank matrix.

Cite this Paper

BibTeX
@InProceedings{pmlr-v235-alder24a, title = {Energy-Efficient {G}aussian Processes Using Low-Precision Arithmetic}, author = {Alder, Nicolas and Herbrich, Ralf}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {955--975}, year = {2024}, editor = {Salakhutdinov, Ruslan and Kolter, Zico and Heller, Katherine and Weller, Adrian and Oliver, Nuria and Scarlett, Jonathan and Berkenkamp, Felix}, volume = {235}, series = {Proceedings of Machine Learning Research}, month = {21--27 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v235/main/assets/alder24a/alder24a.pdf}, url = {https://proceedings.mlr.press/v235/alder24a.html}, abstract = {The widespread use of artificial intelligence requires finding energy-efficient paradigms for the field. We propose to reduce the energy consumption of Gaussian process regression using low-precision floating-point representations. We explore how low-precision representations impact the results of Gaussian process regression and how data set properties, implementation approach, model performance, and energy consumption interact. Our findings show that a well-conditioned kernel matrix allows reducing the energy consumption by up to 89.01% for 98.08% of arithmetic operations with little to no impact on model performance. Our findings are relevant whenever one needs to invert a symmetric full-rank matrix.} }
Endnote
%0 Conference Paper %T Energy-Efficient Gaussian Processes Using Low-Precision Arithmetic %A Nicolas Alder %A Ralf Herbrich %B Proceedings of the 41st International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2024 %E Ruslan Salakhutdinov %E Zico Kolter %E Katherine Heller %E Adrian Weller %E Nuria Oliver %E Jonathan Scarlett %E Felix Berkenkamp %F pmlr-v235-alder24a %I PMLR %P 955--975 %U https://proceedings.mlr.press/v235/alder24a.html %V 235 %X The widespread use of artificial intelligence requires finding energy-efficient paradigms for the field. We propose to reduce the energy consumption of Gaussian process regression using low-precision floating-point representations. We explore how low-precision representations impact the results of Gaussian process regression and how data set properties, implementation approach, model performance, and energy consumption interact. Our findings show that a well-conditioned kernel matrix allows reducing the energy consumption by up to 89.01% for 98.08% of arithmetic operations with little to no impact on model performance. Our findings are relevant whenever one needs to invert a symmetric full-rank matrix.
APA
Alder, N. & Herbrich, R.. (2024). Energy-Efficient Gaussian Processes Using Low-Precision Arithmetic. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:955-975 Available from https://proceedings.mlr.press/v235/alder24a.html.

Related Material