深度学习算法与应用探究

On the Algorithm and Application of Deep Learning

首先对浅层结构经典算法易出现局部最优和过度拟合状况做了分析、归纳.其次,对深度学习现状做了阐述,详细介绍基于RBM和AE算法的深度模型,并对深度学习在智能语音、机器视觉图像、点击通过率、逻辑回归模型、自然语言以及视频动作识别等方面应用作了介绍,表明深度学习相对于其它模型结构优势明显,对深度学习的理论以及应用的研究有重要意义.最后对深度学习当前研究存在的问题进行简要归纳总结,并对模型结构未来研究方向做了探究.

Firstly, this thesis analyzes and summarizes the local optimum and over fitting which are likely to occur in classic surface structure algorithm. Secondly, this thesis illustrates the current situa- tion of deep learning and gives a detailed introduction of deep model based on RBM and AE algorithm. Thirdly, this thesis introduces the application of deep learning in recent years in terms of intelligent voice, machine vision image, click- through rates, logistic regression model, natural language and video motion recognition, etc. and demonstrates the obvious advantages o{ deep learning over other model structures, which has great significance in promoting the research on the theory and application of deep learning. Finally, this thesis summarizes the problems of current research on deep learning briefly and explores the future direction of model structure research.