Boosting for Control of Dynamical Systems

Naman Agarwal, Nataly Brukhim, Elad Hazan, Zhou Lu
Proceedings of the 37th International Conference on Machine Learning, PMLR 119:96-103, 2020.

Abstract

We study the question of how to aggregate controllers for dynamical systems in order to improve their performance. To this end, we propose a framework of boosting for online control. Our main result is an efficient boosting algorithm that combines weak controllers into a provably more accurate one. Empirical evaluation on a host of control settings supports our theoretical findings.

Cite this Paper


BibTeX
@InProceedings{pmlr-v119-agarwal20b, title = {Boosting for Control of Dynamical Systems}, author = {Agarwal, Naman and Brukhim, Nataly and Hazan, Elad and Lu, Zhou}, booktitle = {Proceedings of the 37th International Conference on Machine Learning}, pages = {96--103}, year = {2020}, editor = {III, Hal Daumé and Singh, Aarti}, volume = {119}, series = {Proceedings of Machine Learning Research}, month = {13--18 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v119/agarwal20b/agarwal20b.pdf}, url = {https://proceedings.mlr.press/v119/agarwal20b.html}, abstract = {We study the question of how to aggregate controllers for dynamical systems in order to improve their performance. To this end, we propose a framework of boosting for online control. Our main result is an efficient boosting algorithm that combines weak controllers into a provably more accurate one. Empirical evaluation on a host of control settings supports our theoretical findings.} }
Endnote
%0 Conference Paper %T Boosting for Control of Dynamical Systems %A Naman Agarwal %A Nataly Brukhim %A Elad Hazan %A Zhou Lu %B Proceedings of the 37th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2020 %E Hal Daumé III %E Aarti Singh %F pmlr-v119-agarwal20b %I PMLR %P 96--103 %U https://proceedings.mlr.press/v119/agarwal20b.html %V 119 %X We study the question of how to aggregate controllers for dynamical systems in order to improve their performance. To this end, we propose a framework of boosting for online control. Our main result is an efficient boosting algorithm that combines weak controllers into a provably more accurate one. Empirical evaluation on a host of control settings supports our theoretical findings.
APA
Agarwal, N., Brukhim, N., Hazan, E. & Lu, Z.. (2020). Boosting for Control of Dynamical Systems. Proceedings of the 37th International Conference on Machine Learning, in Proceedings of Machine Learning Research 119:96-103 Available from https://proceedings.mlr.press/v119/agarwal20b.html.

Related Material