arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1503.05431math.NAcs.NA
收藏

交替最小二乘优化在高阶张量秩一近似中的收敛性

Convergence of Alternating Least Squares Optimisation for Rank-One Approximation to High Order Tensors

RWTH Aachen University(亚琛工业大学)

Mike Espig, Aram Khachatryan

AI总结:

研究交替最小二乘法在高阶张量秩一近似中的收敛性与收敛速率,证明该方法可实现亚线性、线性和超线性收敛。

摘要:

张量近似在各学科中有重要应用,但仍是极具挑战性的问题。已知高阶张量可能不存在最佳低秩近似,但最佳秩一近似总是存在的。最流行的低秩近似方法是交替最小二乘法(ALS)。本文分析了ALS算法在秩一近似问题中的收敛性。我们的分析聚焦于ALS算法的全局收敛性和收敛速率。证明ALS方法可实现亚线性、线性甚至超线性收敛。我们的理论结果通过显式例子进行了说明。

英文摘要:

The approximation of tensors has important applications in various disciplines, but it remains an extremely challenging task. It is well known that tensors of higher order can fail to have best low-rank approximations, but with an important exception that best rank-one approximations always exists. The most popular approach to low-rank approximation is the alternating least squares (ALS) method. The convergence of the alternating least squares algorithm for the rank-one approximation problem is analysed in this paper. In our analysis we are focusing on the global convergence and the rate of convergence of the ALS algorithm. It is shown that the ALS method can converge sublinearly, Q-linearly, and even Q-superlinearly. Our theoretical results are illustrated on explicit examples.

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。

↑