【源码】深度神经网络工具箱——提供深度信念网络（DBNs）的深度学习工具

深度神经网络工具箱——提供深度信念网络（DBNs）的深度学习工具

首先，你可以运行testDNN.m尝试一下！

Run testDNN to try!

每个函数都有详细的描述说明。

Each function includes description.

请自己逐一核实查对！

Please check it!

本工具箱提供了堆积有限Boltzmann机（RBMs）的深度信念网络（DBNs）的深度学习工具。

It provides deep learning tools of deepbelief networks (DBNs) of stacked restricted Boltzmann machines (RBMs).

它包括Bernoulli-BernoulliRBM、Gaussian-BernoulliRBM、无监督预训练的对比散度学习、稀疏约束、有监督训练的后向投影以及dropout技术。

It includes the Bernoulli-Bernoulli RBM,the Gaussian-Bernoulli RBM, the contrastive divergence learning forunsupervised pre-training, the sparse constraint, the back projection forsupervised training, and the dropout technique.

MNIST数据集的示例代码包含在MNIST文件夹中。

The sample codes with the MNIST dataset areincluded in the mnist folder.

详细说明请参阅MNIST文件夹中的readme.txt。

Please, see readme.txt in the mnist folder.

参考文献：

Hinton et al, Improving neural networks bypreventing co-adaptation of feature detectors, 2012.

Lee et al, Sparse deep belief net model forvisual area V2, NIPS 2008.

Masayuki Tanaka and Masatoshi Okutomi, ANovel Inference of a Restricted Boltzmann Machine, International Conference onPattern Recognition (ICPR2014), August, 2014.

% randRBM: get randomized restricted boltzmann machine (RBM) model
% 获取随机受限Boltzmann机模型
% rbm = randRBM( dimV, dimH, type )

…
下载源码请点击：

http://page2.dfpan.com/fs/7lc8j2321029916cfc7/

更多精彩文章请关注微信号：