NeurIPS 2021poster4 citations

Reusing Combinatorial Structure: Faster Iterative Projections over Submodular Base Polytopes

Jai Moondra, Hassan Mortagy, Swati Gupta

Abstract

Optimization algorithms such as projected Newton's method, FISTA, mirror descent and its variants enjoy near-optimal regret bounds and convergence rates, but suffer from a computational bottleneck of computing ``projections" in potentially each iteration (e.g., $O(T^{1/2})$ regret of online mirror descent). On the other hand, conditional gradient variants solve a linear optimization in each iteration, but result in suboptimal rates (e.g., $O(T^{3/4})$ regret of online Frank-Wolfe). Motivated by this trade-off in runtime v/s convergence rates, we consider iterative projections of close-by points over widely-prevalent submodular base polytopes $B(f)$. We develop a toolkit to speed up the computation of projections using both discrete and continuous perspectives. We subsequently adapt the away-step Frank-Wolfe algorithm to use this information and enable early termination. For the special case of cardinality based submodular polytopes, we improve the runtime of computing certain Bregman projections by a factor of $\Omega(n/\log(n))$. Our theoretical results show orders of magnitude reduction in runtime in preliminary computational experiments.

Convex optimizationConditional GradientsBregman ProjectionsSubmodular base polytopesOnline learning
BibTeX
@inproceedings{
moondra2021reusing,
title={Reusing Combinatorial Structure: Faster Iterative Projections over Submodular Base Polytopes},
author={Jai Moondra and Hassan Mortagy and Swati Gupta},
booktitle={Advances in Neural Information Processing Systems},
editor={A. Beygelzimer and Y. Dauphin and P. Liang and J. Wortman Vaughan},
year={2021},
url={https://openreview.net/forum?id=961kvwqhR05}
}