Journal ArticleDOI
Some comments of Wolfe's `away step'
J Guélat,Patrice Marcotte +1 more
TLDR
It is given a detailed proof, under slightly weaker conditions on the objective function, that a modified Frank-Wolfe algorithm based on Wolfe's ‘away step’ strategy can achieve geometric convergence, provided a strict complementarity assumption holds.Abstract:
We give a detailed proof, under slightly weaker conditions on the objective function, that a modified Frank-Wolfe algorithm based on Wolfe's ‘away step’ strategy can achieve geometric convergence, provided a strict complementarity assumption holds.read more
Citations
More filters
Journal ArticleDOI
Fast and scalable Lasso via stochastic Frank---Wolfe methods with a convergence guarantee
TL;DR: This paper presents a high-performance implementation of the Frank–Wolfe method tailored to solve large-scale Lasso regression problems, based on a randomized iteration, and proves that the convergence guarantees of the standard FW method are preserved in the stochastic setting.
Posted Content
Sparse Variational Inference: Bayesian Coresets from Scratch
Trevor Campbell,Boyan Beronov +1 more
TL;DR: The proposed Riemannian coreset construction algorithm is fully automated, requiring no problem-specific inputs aside from the probabilistic model and dataset, and able to continually improve the coreset, providing state-of-the-art Bayesian dataset summarization with orders- of-magnitude reduction in KL divergence to the exact posterior.
Proceedings Article
Barrier Frank-Wolfe for marginal inference
TL;DR: In this article, the conditional gradient method (Frank-Wolfe) is used to move pseudomarginals within the marginal polytope through repeated maximum a posteriori (MAP) calls.
Posted Content
A Conditional Gradient Framework for Composite Convex Minimization with Applications to Semidefinite Programming
TL;DR: This approach combines smoothing and homotopy techniques under the CGM framework, and provably achieves the optimal $\mathcal{O}(1/\sqrt{k})$ convergence rate, and demonstrates the convergence when the non-smooth term is an indicator function.
Journal ArticleDOI
Complexity of linear minimization and projection on some sets
TL;DR: A motivation put forward in a large body of work on the Frank-Wolfe algorithm is the computational advantage of solving linear minimizations instead of projections, but the discussions supporting this advantage are often incomplete.
References
More filters