Communication-Efficient Distributed PCA by Riemannian Optimization

Communication-Efficient Distributed PCA by Riemannian Optimization
复制标题

通过黎曼优化实现通信高效的分布式 PCA

DOI:
--
复制
发表时间:
2020
期刊:
International Conference on Machine Learning
影响因子:
--
通讯作者:
Sinno Jialin Pan
Sinno Jialin Pan
中科院分区:
--
文献类型:
--
作者:
Long;Sinno Jialin Pan;Sinno Jialin Pan

文献摘要

参考文献

被引文献

相似文献

在本文中,我们研究了统计分布设置的领先特征向量问题,并提出了一个通信效率的算法的基础上黎曼优化,贸易的局部计算的全球通信。理论分析表明,所提出的算法线性收敛到集中的经验风险最小化的解决方案的通信轮数。当本地机器中的数据点的数量足够大时,所提出的算法实现了显着降低通信成本比现有的分布式PCA算法。在真实世界和合成数据集上验证了所提出的算法在通信成本方面的上级性能。
In this paper, we study the leading eigenvector problem in a statistically distributed setting and propose a communication-efficient algorithm based on Riemannian optimization, which trades local computation for global communication. Theoretical analysis shows that the proposed algorithm linearly converges to the centralized empirical risk minimization solution regarding the number of communication rounds. When the number of data points in local machines is sufficiently large, the proposed algorithm achieves a significant reduction of communication cost over existing distributed PCA algorithms. Superior performance in terms of communication cost of the proposed algorithm is verified on real-world and synthetic datasets.
DOI: --
发表时间: 2018
期刊: Advances in Neural Information Processing Systems 31
影响因子: --
作者:
Saparbayeva, Bayan;Zhang, Michael;Lin, Lizhen
通讯作者: Lin, Lizhen
DOI: 10.1080/01621459.2018.1429274
发表时间: 2019-04-03
影响因子: 3.7
作者:
Jordan, Michael I.;Lee, Jason D.;Yang, Yun
通讯作者: Yang, Yun