Projection-Free Online Optimization with Stochastic Gradient: From Convexity to Submodularity

Lin Chen, Christopher Harshaw, Hamed Hassani, Amin Karbasi
Proceedings of the 35th International Conference on Machine Learning, PMLR 80:814-823, 2018.

Abstract

Online optimization has been a successful framework for solving large-scale problems under computational constraints and partial information. Current methods for online convex optimization require either a projection or exact gradient computation at each step, both of which can be prohibitively expensive for large-scale applications. At the same time, there is a growing trend of non-convex optimization in machine learning community and a need for online methods. Continuous DR-submodular functions, which exhibit a natural diminishing returns condition, have recently been proposed as a broad class of non-convex functions which may be efficiently optimized. Although online methods have been introduced, they suffer from similar problems. In this work, we propose Meta-Frank-Wolfe, the first online projection-free algorithm that uses stochastic gradient estimates. The algorithm relies on a careful sampling of gradients in each round and achieves the optimal $O( \sqrt{T})$ adversarial regret bounds for convex and continuous submodular optimization. We also propose One-Shot Frank-Wolfe, a simpler algorithm which requires only a single stochastic gradient estimate in each round and achieves an $O(T^{2/3})$ stochastic regret bound for convex and continuous submodular optimization. We apply our methods to develop a novel "lifting" framework for the online discrete submodular maximization and also see that they outperform current state-of-the-art techniques on various experiments.

Cite this Paper


BibTeX
@InProceedings{pmlr-v80-chen18c, title = {Projection-Free Online Optimization with Stochastic Gradient: From Convexity to Submodularity}, author = {Chen, Lin and Harshaw, Christopher and Hassani, Hamed and Karbasi, Amin}, booktitle = {Proceedings of the 35th International Conference on Machine Learning}, pages = {814--823}, year = {2018}, editor = {Dy, Jennifer and Krause, Andreas}, volume = {80}, series = {Proceedings of Machine Learning Research}, month = {10--15 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v80/chen18c/chen18c.pdf}, url = {https://proceedings.mlr.press/v80/chen18c.html}, abstract = {Online optimization has been a successful framework for solving large-scale problems under computational constraints and partial information. Current methods for online convex optimization require either a projection or exact gradient computation at each step, both of which can be prohibitively expensive for large-scale applications. At the same time, there is a growing trend of non-convex optimization in machine learning community and a need for online methods. Continuous DR-submodular functions, which exhibit a natural diminishing returns condition, have recently been proposed as a broad class of non-convex functions which may be efficiently optimized. Although online methods have been introduced, they suffer from similar problems. In this work, we propose Meta-Frank-Wolfe, the first online projection-free algorithm that uses stochastic gradient estimates. The algorithm relies on a careful sampling of gradients in each round and achieves the optimal $O( \sqrt{T})$ adversarial regret bounds for convex and continuous submodular optimization. We also propose One-Shot Frank-Wolfe, a simpler algorithm which requires only a single stochastic gradient estimate in each round and achieves an $O(T^{2/3})$ stochastic regret bound for convex and continuous submodular optimization. We apply our methods to develop a novel "lifting" framework for the online discrete submodular maximization and also see that they outperform current state-of-the-art techniques on various experiments.} }
Endnote
%0 Conference Paper %T Projection-Free Online Optimization with Stochastic Gradient: From Convexity to Submodularity %A Lin Chen %A Christopher Harshaw %A Hamed Hassani %A Amin Karbasi %B Proceedings of the 35th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2018 %E Jennifer Dy %E Andreas Krause %F pmlr-v80-chen18c %I PMLR %P 814--823 %U https://proceedings.mlr.press/v80/chen18c.html %V 80 %X Online optimization has been a successful framework for solving large-scale problems under computational constraints and partial information. Current methods for online convex optimization require either a projection or exact gradient computation at each step, both of which can be prohibitively expensive for large-scale applications. At the same time, there is a growing trend of non-convex optimization in machine learning community and a need for online methods. Continuous DR-submodular functions, which exhibit a natural diminishing returns condition, have recently been proposed as a broad class of non-convex functions which may be efficiently optimized. Although online methods have been introduced, they suffer from similar problems. In this work, we propose Meta-Frank-Wolfe, the first online projection-free algorithm that uses stochastic gradient estimates. The algorithm relies on a careful sampling of gradients in each round and achieves the optimal $O( \sqrt{T})$ adversarial regret bounds for convex and continuous submodular optimization. We also propose One-Shot Frank-Wolfe, a simpler algorithm which requires only a single stochastic gradient estimate in each round and achieves an $O(T^{2/3})$ stochastic regret bound for convex and continuous submodular optimization. We apply our methods to develop a novel "lifting" framework for the online discrete submodular maximization and also see that they outperform current state-of-the-art techniques on various experiments.
APA
Chen, L., Harshaw, C., Hassani, H. & Karbasi, A.. (2018). Projection-Free Online Optimization with Stochastic Gradient: From Convexity to Submodularity. Proceedings of the 35th International Conference on Machine Learning, in Proceedings of Machine Learning Research 80:814-823 Available from https://proceedings.mlr.press/v80/chen18c.html.

Related Material