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
Posted Content
Frank-Wolfe and friends: a journey into projection-free first-order optimization methods.
TL;DR: In this paper, a review of the success of the Frank-Wolfe method in a wide range of contexts is presented, together with an account on recent progress in variants, both improving on the speed and efficiency of first-order optimization.
Posted Content
Discovering a set of policies for the worst case reward
Tom Zahavy,Andre Barreto,Daniel J. Mankowitz,Shaobo Hou,Brendan O'Donoghue,Iurii Kemaev,Satinder Singh +6 more
TL;DR: In this paper, a policy iteration algorithm is proposed to construct a set of policies in order to maximize the worst-case performance of the resulting SMP on the set of tasks, and the algorithm only stops when there does not exist a policy that leads to improved performance.
Posted Content
Frank-Wolfe Algorithm for the Exact Sparse Problem
TL;DR: In this paper, the authors studied the properties of the Frank-Wolfe algorithm to solve the exact sparse reconstruction problem and showed that when the dictionary is quasi-incoherent, at each iteration, the Frank Wolfe algorithm picks up an atom indexed by the support.
Journal Article
Compact convex projections
TL;DR: This work derives non-parametric regression algorithms which use at their core a slightly modified kernel herding algorithm to determine projections and demonstrates via experiments that the developed regressors are en-par with state-of-the-art regression algorithms for large scale problems.
A Cyclic Coordinate Descent Method for Convex Optimization on Polytopes
Rahul Mazumder,Haoyue Wang +1 more
TL;DR: PolyCDwA as mentioned in this paper is a hybrid of cyclic coordinate descent and Frank-Wolfe algorithm with away steps, which has a linear convergence rate when the loss function is smooth and strongly convex.
References
More filters