Making Frank-Wolfe algorithms practical and scalable.
We propose a novel Stochastic Frank-Wolfe (a.k.a. Conditional Gradient) algorithm with a fixed batch size tailored to the constrained optimization of a finite sum of smooth objectives. The design of our method hinges on a primal-dual interpretation …
Structured constraints in Machine Learning have recently brought the Frank-Wolfe (FW) family of algorithms back in the spotlight. Recently, the Away-steps (A) and Pairwise (P) FW variants have been shown to converge linearly for polytopic …
Structured constraints in Machine Learning have recently brought the Frank-Wolfe (FW) family of algorithms back in the spotlight. While the classical FW algorithm has poor local convergence properties, the Away-steps and Pairwise FW variants have …