自动化学报2017，Vol.43Issue(8)：1339-1349,11.DOI:10.16383/j.aas.2017.c160389
基于自适应学习率的深度信念网设计与应用
Design and Application of Deep Belief Network with Adaptive Learning Rate
摘要
Abstract
A deep belief network with adaptive learning rate (ALRDBN) is proposed to solve the time-consuming problem in the pre-training period of DBN. The ALRDBN introduces the idea of adaptive learning rate into contrastive divergence (CD) algorithm and accelerates its convergence by a self-adjusting learning rate. The training method of weights in this case is designed, in which the adjusting scope of the coefficient in learning rate is determined by performance analysis. Finally, a series of experiments are carried out to test the performance of ALRDBN, and the corresponding results show that the convergence rate is accelerated significantly and the accuracy of prediction is improved as well.关键词
深度信念网/自适应学习率/对比差度/收敛速度/性能分析Key words
Deep belief network/adaptive learning rate/contrastive divergence/convergence rate/performance analysis引用本文复制引用
乔俊飞,王功明,李晓理,韩红桂,柴伟..基于自适应学习率的深度信念网设计与应用[J].自动化学报,2017,43(8):1339-1349,11.基金项目
国家自然科学基金(61533002, 61473034), 国家杰出青年科学基金(61225016), 内涵发展—引进人才科研启动费资助Supported by National Natural Science Foundation of China (61533002, 61473034), National Natural Science Fund for Distin-guished Young Scholars (61225016), Connotation Development—Scientific Research Start-up Funds of Talent Introduction （61533002, 61473034）
