Applied Online Algorithms with Heterogeneous Predictors

Jessica Maghakian, Russell Lee, Mohammad Hajiesmaili, Jian Li, Ramesh Sitaraman, Zhenhua Liu
Proceedings of the 40th International Conference on Machine Learning, PMLR 202:23484-23497, 2023.

Abstract

For many application domains, the integration of machine learning (ML) models into decision making is hindered by the poor explainability and theoretical guarantees of black box models. Although the emerging area of algorithms with predictions offers a way to leverage ML while enjoying worst-case guarantees, existing work usually assumes access to only one predictor. We demonstrate how to more effectively utilize historical datasets and application domain knowledge by intentionally using predictors of different quantities. By leveraging the heterogeneity in our predictors, we are able to achieve improved performance, explainability and computational efficiency over predictor-agnostic methods. Theoretical results are supplemented by large-scale empirical evaluations with production data demonstrating the success of our methods on optimization problems occurring in large distributed computing systems.

Cite this Paper

BibTeX
@InProceedings{pmlr-v202-maghakian23a, title = {Applied Online Algorithms with Heterogeneous Predictors}, author = {Maghakian, Jessica and Lee, Russell and Hajiesmaili, Mohammad and Li, Jian and Sitaraman, Ramesh and Liu, Zhenhua}, booktitle = {Proceedings of the 40th International Conference on Machine Learning}, pages = {23484--23497}, year = {2023}, editor = {Krause, Andreas and Brunskill, Emma and Cho, Kyunghyun and Engelhardt, Barbara and Sabato, Sivan and Scarlett, Jonathan}, volume = {202}, series = {Proceedings of Machine Learning Research}, month = {23--29 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v202/maghakian23a/maghakian23a.pdf}, url = {https://proceedings.mlr.press/v202/maghakian23a.html}, abstract = {For many application domains, the integration of machine learning (ML) models into decision making is hindered by the poor explainability and theoretical guarantees of black box models. Although the emerging area of algorithms with predictions offers a way to leverage ML while enjoying worst-case guarantees, existing work usually assumes access to only one predictor. We demonstrate how to more effectively utilize historical datasets and application domain knowledge by intentionally using predictors of different quantities. By leveraging the heterogeneity in our predictors, we are able to achieve improved performance, explainability and computational efficiency over predictor-agnostic methods. Theoretical results are supplemented by large-scale empirical evaluations with production data demonstrating the success of our methods on optimization problems occurring in large distributed computing systems.} }
Endnote
%0 Conference Paper %T Applied Online Algorithms with Heterogeneous Predictors %A Jessica Maghakian %A Russell Lee %A Mohammad Hajiesmaili %A Jian Li %A Ramesh Sitaraman %A Zhenhua Liu %B Proceedings of the 40th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2023 %E Andreas Krause %E Emma Brunskill %E Kyunghyun Cho %E Barbara Engelhardt %E Sivan Sabato %E Jonathan Scarlett %F pmlr-v202-maghakian23a %I PMLR %P 23484--23497 %U https://proceedings.mlr.press/v202/maghakian23a.html %V 202 %X For many application domains, the integration of machine learning (ML) models into decision making is hindered by the poor explainability and theoretical guarantees of black box models. Although the emerging area of algorithms with predictions offers a way to leverage ML while enjoying worst-case guarantees, existing work usually assumes access to only one predictor. We demonstrate how to more effectively utilize historical datasets and application domain knowledge by intentionally using predictors of different quantities. By leveraging the heterogeneity in our predictors, we are able to achieve improved performance, explainability and computational efficiency over predictor-agnostic methods. Theoretical results are supplemented by large-scale empirical evaluations with production data demonstrating the success of our methods on optimization problems occurring in large distributed computing systems.
APA
Maghakian, J., Lee, R., Hajiesmaili, M., Li, J., Sitaraman, R. & Liu, Z.. (2023). Applied Online Algorithms with Heterogeneous Predictors. Proceedings of the 40th International Conference on Machine Learning, in Proceedings of Machine Learning Research 202:23484-23497 Available from https://proceedings.mlr.press/v202/maghakian23a.html.

Related Material