Revisiting Frank-Wolfe: Projection-Free Sparse Convex Optimization

Martin Jaggi
Proceedings of the 30th International Conference on Machine Learning, PMLR 28(1):427-435, 2013.

Abstract

We provide stronger and more general primal-dual convergence results for Frank-Wolfe-type algorithms (a.k.a. conditional gradient) for constrained convex optimization, enabled by a simple framework of duality gap certificates. Our analysis also holds if the linear subproblems are only solved approximately (as well as if the gradients are inexact), and is proven to be worst-case optimal in the sparsity of the obtained solutions. On the application side, this allows us to unify a large variety of existing sparse greedy methods, in particular for optimization over convex hulls of an atomic set, even if those sets can only be approximated, including sparse (or structured sparse) vectors or matrices, low-rank matrices, permutation matrices, or max-norm bounded matrices. We present a new general framework for convex optimization over matrix factorizations, where every Frank-Wolfe iteration will consist of a low-rank update, and discuss the broad application areas of this approach.

Cite this Paper


BibTeX
@InProceedings{pmlr-v28-jaggi13, title = {Revisiting {Frank-Wolfe}: Projection-Free Sparse Convex Optimization}, author = {Jaggi, Martin}, booktitle = {Proceedings of the 30th International Conference on Machine Learning}, pages = {427--435}, year = {2013}, editor = {Dasgupta, Sanjoy and McAllester, David}, volume = {28}, number = {1}, series = {Proceedings of Machine Learning Research}, address = {Atlanta, Georgia, USA}, month = {17--19 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v28/jaggi13.pdf}, url = {https://proceedings.mlr.press/v28/jaggi13.html}, abstract = {We provide stronger and more general primal-dual convergence results for Frank-Wolfe-type algorithms (a.k.a. conditional gradient) for constrained convex optimization, enabled by a simple framework of duality gap certificates. Our analysis also holds if the linear subproblems are only solved approximately (as well as if the gradients are inexact), and is proven to be worst-case optimal in the sparsity of the obtained solutions. On the application side, this allows us to unify a large variety of existing sparse greedy methods, in particular for optimization over convex hulls of an atomic set, even if those sets can only be approximated, including sparse (or structured sparse) vectors or matrices, low-rank matrices, permutation matrices, or max-norm bounded matrices. We present a new general framework for convex optimization over matrix factorizations, where every Frank-Wolfe iteration will consist of a low-rank update, and discuss the broad application areas of this approach.} }
Endnote
%0 Conference Paper %T Revisiting Frank-Wolfe: Projection-Free Sparse Convex Optimization %A Martin Jaggi %B Proceedings of the 30th International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2013 %E Sanjoy Dasgupta %E David McAllester %F pmlr-v28-jaggi13 %I PMLR %P 427--435 %U https://proceedings.mlr.press/v28/jaggi13.html %V 28 %N 1 %X We provide stronger and more general primal-dual convergence results for Frank-Wolfe-type algorithms (a.k.a. conditional gradient) for constrained convex optimization, enabled by a simple framework of duality gap certificates. Our analysis also holds if the linear subproblems are only solved approximately (as well as if the gradients are inexact), and is proven to be worst-case optimal in the sparsity of the obtained solutions. On the application side, this allows us to unify a large variety of existing sparse greedy methods, in particular for optimization over convex hulls of an atomic set, even if those sets can only be approximated, including sparse (or structured sparse) vectors or matrices, low-rank matrices, permutation matrices, or max-norm bounded matrices. We present a new general framework for convex optimization over matrix factorizations, where every Frank-Wolfe iteration will consist of a low-rank update, and discuss the broad application areas of this approach.
RIS
TY - CPAPER TI - Revisiting Frank-Wolfe: Projection-Free Sparse Convex Optimization AU - Martin Jaggi BT - Proceedings of the 30th International Conference on Machine Learning DA - 2013/02/13 ED - Sanjoy Dasgupta ED - David McAllester ID - pmlr-v28-jaggi13 PB - PMLR DP - Proceedings of Machine Learning Research VL - 28 IS - 1 SP - 427 EP - 435 L1 - http://proceedings.mlr.press/v28/jaggi13.pdf UR - https://proceedings.mlr.press/v28/jaggi13.html AB - We provide stronger and more general primal-dual convergence results for Frank-Wolfe-type algorithms (a.k.a. conditional gradient) for constrained convex optimization, enabled by a simple framework of duality gap certificates. Our analysis also holds if the linear subproblems are only solved approximately (as well as if the gradients are inexact), and is proven to be worst-case optimal in the sparsity of the obtained solutions. On the application side, this allows us to unify a large variety of existing sparse greedy methods, in particular for optimization over convex hulls of an atomic set, even if those sets can only be approximated, including sparse (or structured sparse) vectors or matrices, low-rank matrices, permutation matrices, or max-norm bounded matrices. We present a new general framework for convex optimization over matrix factorizations, where every Frank-Wolfe iteration will consist of a low-rank update, and discuss the broad application areas of this approach. ER -
APA
Jaggi, M.. (2013). Revisiting Frank-Wolfe: Projection-Free Sparse Convex Optimization. Proceedings of the 30th International Conference on Machine Learning, in Proceedings of Machine Learning Research 28(1):427-435 Available from https://proceedings.mlr.press/v28/jaggi13.html.

Related Material