arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2006.16142math.OCcs.LG
收藏

$k$ FW:一个具有更强子问题预言的Frank-Wolfe风格算法

$k$FW: A Frank-Wolfe style algorithm with stronger subproblem oracles

The Chinese University of Hong Kong(香港中文大学);Cornell University(康奈尔大学)

Lijun Ding, Jicong Fan, Madeleine Udell

摘要:

提出了一种新的Frank-Wolfe(FW)算法,称为$k$ FW。标准FW收敛速度慢:迭代通常是Z字形的,因为更新方向围绕约束集的极值点振荡。新的变体,$k$ FW,克服了这个问题,在每次迭代中使用两个更强的子问题预言。第一个是计算$k$最佳更新方向(而不是仅一个)的$k$线性最优化oracle($k$ LOO)。第二个是$k$方向搜索($k$ DS),其在由$k$最佳更新方向和先前的方向表示的约束集上最小化目标。当问题的解决方案承认一个稀疏的表示,这两个预言是很容易计算,和$k$ FW收敛迅速光滑凸目标和几个有趣的约束集:$k$ FW实现有限的$\frac{4L_f^3D^4}{γδ^2}$收敛多面体和群范数球,和线性收敛的谱面体和核范数球。数值实验验证了$k$ FW的有效性,并证明了一个数量级的速度比现有的方法。

英文摘要:

This paper proposes a new variant of Frank-Wolfe (FW), called $k$FW. Standard FW suffers from slow convergence: iterates often zig-zag as update directions oscillate around extreme points of the constraint set. The new variant, $k$FW, overcomes this problem by using two stronger subproblem oracles in each iteration. The first is a $k$ linear optimization oracle ($k$LOO) that computes the $k$ best update directions (rather than just one). The second is a $k$ direction search ($k$DS) that minimizes the objective over a constraint set represented by the $k$ best update directions and the previous iterate. When the problem solution admits a sparse representation, both oracles are easy to compute, and $k$FW converges quickly for smooth convex objectives and several interesting constraint sets: $k$FW achieves finite $\frac{4L_f^3D^4}{γδ^2}$ convergence on polytopes and group norm balls, and linear convergence on spectrahedra and nuclear norm balls. Numerical experiments validate the effectiveness of $k$FW and demonstrate an order-of-magnitude speedup over existing approaches.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑