OCTEN: Online Compression-Based Tensor Decomposition

OCTEN: Online Compression-Based Tensor Decomposition
复制标题

OCTEN:在线基于压缩的张量分解

DOI:
10.1109/camsap45676.2019.9022641
复制
发表时间:
2019
期刊:
2019 IEEE 8th International Workshop on Computational Advances in Multi-Sensor Adaptive Processing (CAMSAP
影响因子:
--
通讯作者:
Papalexakis, Evangelos E.
Papalexakis, Evangelos E.
中科院分区:
--
文献类型:
--
作者:
Gujral, Ekta;Pasricha, Ravdeep;Yang, Tianxiong;Papalexakis, Evangelos E.

文献摘要

参考文献

被引文献

相似文献

张量分解是大数据分析的强大工具,因为它们将数据的多个方面联合建模到一个框架中,并能够发现数据中的潜在结构和高阶相关性。最广泛研究和使用的分解之一,特别是在数据挖掘和机器学习中,是正则多元分解或 PARAFAC 分解。然而,当今的数据集并不是静态的,并且经常随着时间的推移而增长和变化。为了处理如此大的动态数据,我们提出了 OCTEN,这是第一个基于压缩的 CP/PARAFAC 分解在线并行实现。我们对算法的适应度、内存使用和 CPU 时间进行了广泛的实证分析,为了证明该方法的压缩和可扩展性,我们将 OCTEN 应用于大张量数据。可以看出,OCTEN 在分解精度和效率方面与最先进的在线和离线方法相当甚至更好,同时节省了 40-200% 的内存。
Tensor decompositions are powerful tools for large data analytics, as they jointly model multiple aspects of data into one framework and enable the discovery of the latent structures and higher-order correlations within the data. One of the most widely studied and used decompositions, especially in data mining and machine learning, is the Canonical Polyadic or PARAFAC decomposition. However, today's datasets are not static and often grow and change over time. To operate on such large dynamic data, we present OCTEN, the first ever compression-based online parallel implementation for the CP/PARAFAC decomposition. We conduct an extensive empirical analysis of the algorithms in terms of fitness, memory used and CPU time and in order to demonstrate the compression and scalability of the method, we apply OCTEN to big tensor data. Indicatively, OCTEN performs on-par or better than state-of-the-art online and offline methods in terms of decomposition accuracy and efficiency, while achieving memory savings ranging in 40-200%.
DOI: 10.1137/1.9781611975321.44
发表时间: 2017-09
期刊: ArXiv
影响因子: --
作者:
Ekta Gujral;Ravdeep Pasricha;E. Papalexakis
通讯作者: Ekta Gujral;Ravdeep Pasricha;E. Papalexakis
DOI: 10.1109/tsmc.2014.2327053
发表时间: 2015-01-01
影响因子: 8.7
作者:
Yang, Dingqi;Zhang, Daqing;Yu, Zhiyong
通讯作者: Yu, Zhiyong