GPU加速的神经网络BP算法  

GPU accelerated BP algorithm of neural network

作　　者：田绪红[1] 江敏杰[1] 

机构地区：[1]华南农业大学信息学院,广州510642

出　　处：《计算机应用研究》2009年第5期1679-1681,1699,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(60573019);广东省科技计划基金资助项目(2007B020706006;2007B020715001)

摘　　要：近年来图形处理器(GPU)快速拓展的可编程性能力加上渲染流水线的高速度及并行性,使得图形处理器通用计算(GPGPU)迅速成为一个研究热点。针对大规模神经网络BP算法效率低下问题,提出了一种GPU加速的神经网络BP算法。将BP网络的前向计算、反向学习转换为GPU纹理的渲染过程,从而利用GPU强大的浮点运算能力和高度并行的计算特性对BP算法进行求解。实验结果表明,在保证求解结果准确度不变的情况下,该方法运行效率有明显的提高。Recently, the raw speed, highly data-parallel nature and rapidly expanding programmability of graphics processing units make them an attractive platform for general purpose computation. As the efficiency of BP algorithm in large-scale neural network is relative low, proposed a GPU accelerated BP algorithm of neural network. Translated the forward computing and back-propagation learning of BP algorithm into texture rendering of GPU so as to solve the BP algorithm using the powerful float-point operation ability and high parallel computing characteristic of GPU. Experimental results show that the proposed method greatly raises the speed of BP algorithm in large-scale neural networks without losing the accuracy.

关 键 词：图形处理器 神经网络 反向传播算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]