深度学习算法实验平台的研究与实现.docxVIP

武汉理工大学毕业设计（论文）  DOCPROPERTY PO_TopicName \* MERGEFORMAT 深度学习算法实验平台的研究与实现 学院（系）： 自动化学院 专业班级： 自动化专业1104班 姓名： 指导教师： 摘要 深度学习是基于人工神经网络的机器学习，机器学习方法也有监督学习与无监督学习之分。不同的学习框架下建立的学习模型很不同．例如，卷积神经网络就是一种监督学习下的机器学习模型，而深度置信网就是一种无监督学习下的机器学习模型。 MATLAB作为一款功能强大的科学计算软件，广泛应用于机器学习、图像绘制等许多方面。MATLAB常常作为研究神经网络的入门平台，利用其方便快捷的MATLAB编程语言，能够更好的实现神经网络的构建。作为另一款常用的深度学习平台，python则以其免费的特性受到亲睐，众多开源的科学计算库都提供了Python的调用接口。用户可以在任何计算机上免费安装Python及其绝大多数扩展库。 以MATLAB和Python作为研究平台，实现深度学习算法。比较它们的优缺点，在此基础上完成毕业论文的撰写和完善。 关键词：深度学习;MATLAB;Python;机器学习 Abstract Depth learning is based on artificial neural network machine study, machine learning methods have supervised learning and unsupervised learning points. Learning models at different learning framework established very different. For example, convolution neural network is a supervised learning machine learning models under, and the depth of belief network is a non-supervised machine learning models under study. MATLAB as a powerful scientific computing software, widely used in machine learning, image rendering, and many other aspects. MATLAB neural network is often used as entry platform, using its convenient MATLAB programming language, can better achieve build neural networks. Another commonly used as a deep learning platform, python is its features for free by pro-gaze, many open source scientific computing libraries provide Python call interface. Users are free to install Python and the vast majority of extensions on any computer. With MATLAB and Python as a research platform, deep learning algorithms. Compare their advantages and disadvantages, on the basis of complete and perfect thesis writing. Key Word:Deep learning;MATLAB;Python;unsupervised learning  目录TOC \o 1-3 \h \u   HYPERLINK \l _Toc23120 第1章 绪论  PAGEREF _Toc23120 1  HYPERLINK \l _Toc30178 1.1课题研究意义  PAGEREF _Toc30178 1  HYPERLINK \l _Toc7598 1.2深度学习的基本思想  P