Online Algorithms with Uncertainty-Quantified Predictions

Bo Sun, Jerry Huang, Nicolas Christianson, Mohammad Hajiesmaili, Adam Wierman, Raouf Boutaba
Proceedings of the 41st International Conference on Machine Learning, PMLR 235:47056-47077, 2024.

Abstract

The burgeoning field of algorithms with predictions studies the problem of using possibly imperfect machine learning predictions to improve online algorithm performance. While nearly all existing algorithms in this framework make no assumptions on prediction quality, a number of methods providing uncertainty quantification (UQ) on machine learning models have been developed in recent years, which could enable additional information about prediction quality at decision time. In this work, we investigate the problem of optimally utilizing uncertainty-quantified predictions in the design of online algorithms. In particular, we study two classic online problems, ski rental and online search, where the decision-maker is provided predictions augmented with UQ describing the likelihood of the ground truth falling within a particular range of values. We demonstrate that non-trivial modifications to algorithm design are needed to fully leverage the UQ predictions. Moreover, we consider how to utilize more general forms of UQ, proposing an online learning framework that learns to exploit UQ to make decisions in multi-instance settings.

Cite this Paper


BibTeX
@InProceedings{pmlr-v235-sun24f, title = {Online Algorithms with Uncertainty-Quantified Predictions}, author = {Sun, Bo and Huang, Jerry and Christianson, Nicolas and Hajiesmaili, Mohammad and Wierman, Adam and Boutaba, Raouf}, booktitle = {Proceedings of the 41st International Conference on Machine Learning}, pages = {47056--47077}, 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/sun24f/sun24f.pdf}, url = {https://proceedings.mlr.press/v235/sun24f.html}, abstract = {The burgeoning field of algorithms with predictions studies the problem of using possibly imperfect machine learning predictions to improve online algorithm performance. While nearly all existing algorithms in this framework make no assumptions on prediction quality, a number of methods providing uncertainty quantification (UQ) on machine learning models have been developed in recent years, which could enable additional information about prediction quality at decision time. In this work, we investigate the problem of optimally utilizing uncertainty-quantified predictions in the design of online algorithms. In particular, we study two classic online problems, ski rental and online search, where the decision-maker is provided predictions augmented with UQ describing the likelihood of the ground truth falling within a particular range of values. We demonstrate that non-trivial modifications to algorithm design are needed to fully leverage the UQ predictions. Moreover, we consider how to utilize more general forms of UQ, proposing an online learning framework that learns to exploit UQ to make decisions in multi-instance settings.} }
Endnote
%0 Conference Paper %T Online Algorithms with Uncertainty-Quantified Predictions %A Bo Sun %A Jerry Huang %A Nicolas Christianson %A Mohammad Hajiesmaili %A Adam Wierman %A Raouf Boutaba %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-sun24f %I PMLR %P 47056--47077 %U https://proceedings.mlr.press/v235/sun24f.html %V 235 %X The burgeoning field of algorithms with predictions studies the problem of using possibly imperfect machine learning predictions to improve online algorithm performance. While nearly all existing algorithms in this framework make no assumptions on prediction quality, a number of methods providing uncertainty quantification (UQ) on machine learning models have been developed in recent years, which could enable additional information about prediction quality at decision time. In this work, we investigate the problem of optimally utilizing uncertainty-quantified predictions in the design of online algorithms. In particular, we study two classic online problems, ski rental and online search, where the decision-maker is provided predictions augmented with UQ describing the likelihood of the ground truth falling within a particular range of values. We demonstrate that non-trivial modifications to algorithm design are needed to fully leverage the UQ predictions. Moreover, we consider how to utilize more general forms of UQ, proposing an online learning framework that learns to exploit UQ to make decisions in multi-instance settings.
APA
Sun, B., Huang, J., Christianson, N., Hajiesmaili, M., Wierman, A. & Boutaba, R.. (2024). Online Algorithms with Uncertainty-Quantified Predictions. Proceedings of the 41st International Conference on Machine Learning, in Proceedings of Machine Learning Research 235:47056-47077 Available from https://proceedings.mlr.press/v235/sun24f.html.

Related Material