本文已被：浏览 2136次 下载 4117次
Received:December 28, 2015 Revised:January 27, 2016
Received:December 28, 2015 Revised:January 27, 2016
中文摘要: 深度学习是近年来机器学习的研究热点，并已广泛应用于不同领域. 但由于训练模型复杂和训练集规模庞大等原因导致的深度学习性能问题已成为其发展的一大阻碍. 近年来计算机硬件的快速发展，尤其是处理器核数的不断增加和整体运算能力的快速提高，给深度学习加速提供了硬件基础，然而其训练算法并行度低和内存开销巨大等问题使得加速研究工作困难重重. 首先介绍了深度学习的背景和训练算法，对当前主要的深度学习加速研究工作进行归纳总结. 在此基础上，对经典的深度学习模型进行性能测试，分析了深度学习及并行算法的性能问题. 最后，对深度学习的未来发展进行了展望.
Abstract:Deep learning has become a popular research direction recently and has been applied to many fields. But the performance issue caused by the complicated learning model and massive training data has becomes an obstacle of deep learning evolution. As the development of processor technology, the core number and performance of the processor have increased rapidly. However, the acceleration research is restricted mainly due to the low parallelism and high memory consumption of the training algorithm. The paper introduces the background of deep learning and its training algorithm, then summarizes current deep learning acceleration works. Further, it analyzes classic deep models training algorithm to explain the causes of deep learning performance problem. Based on the analysis, it lists the challenges about deep learning evolution and makes proposal to address them.
keywords: deep learning deep learning acceleration deep learning performance distribution deep learning computing hardware acceleration
文章编号： 中图分类号： 文献标志码：
基金项目:国家高技术研究发展计划（863）（2012AA010905）；国家自然科学基金（61370081）
引用文本：
杨旭瑜,张铮,张为华.深度学习加速技术研究.计算机系统应用,2016,25(9):1-9
YANG Xu-Yu,ZHANG Zheng,ZHANG Wei-Hua.Research on Deep Learning Acceleration Technique.COMPUTER SYSTEMS APPLICATIONS,2016,25(9):1-9
杨旭瑜,张铮,张为华.深度学习加速技术研究.计算机系统应用,2016,25(9):1-9
YANG Xu-Yu,ZHANG Zheng,ZHANG Wei-Hua.Research on Deep Learning Acceleration Technique.COMPUTER SYSTEMS APPLICATIONS,2016,25(9):1-9
