arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1504.01169stat.MLcs.LG
收藏

基于非常稀疏随机投影的高效字典学习

Efficient Dictionary Learning via Very Sparse Random Projections

University of Colorado at Boulder(科罗拉多大学博尔德分校)

Farhad Pourkamali-Anaraki, Stephen Becker, Shannon M. Hughes

摘要:

对数据的压缩测量执行信号处理任务近年来受到极大关注。在本文中,我们通过展示可以使用更一般的随机投影(包括稀疏投影)来扩展先前关于压缩字典学习的工作。更确切地说,我们研究压缩K-means聚类作为压缩字典学习的一个特殊情况,并给出了理论保证其性能的一个非常一般的随机投影类。然后,我们提出了一个内存和计算效率的字典学习算法,专门设计用于分析大量的高维数据,学习字典从非常稀疏的随机投影。实验结果表明,我们的方法允许减少计算复杂度和内存/数据访问,与可控的损失的准确性。

英文摘要:

Performing signal processing tasks on compressive measurements of data has received great attention in recent years. In this paper, we extend previous work on compressive dictionary learning by showing that more general random projections may be used, including sparse ones. More precisely, we examine compressive K-means clustering as a special case of compressive dictionary learning and give theoretical guarantees for its performance for a very general class of random projections. We then propose a memory and computation efficient dictionary learning algorithm, specifically designed for analyzing large volumes of high-dimensional data, which learns the dictionary from very sparse random projections. Experimental results demonstrate that our approach allows for reduction of computational complexity and memory/data access, with controllable loss in accuracy.

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。

↑