School of Computer Science and Technology，China University of Mining and Technology，Xuzhou，221116，China
{{custom_authorNodes}}
{{custom_bio.content}}
{{custom_bio.content}}
{{custom_authorNodes}}
Collapse
History+
Published
2017-05-30
Issue Date
2017-05-30
Abstract
Deep learning is currently an extremely active research area in machine learning and pattern recognition society.Deep learning and Big Data are two hottest trends in rapidly growing digital world.Today，big data for the development of various industry has brought the huge opportunity and potential.On the other hand it has also brought unprecedented challenges.While big data has been defined in different ways，it is referred to the exponential growth and wide availability of digital data.It is difficult or even impossible to manage and analyze them using conventional software tools and technologies.It has gained huge successes in a broad area of applications such as speech recognition，computer vision and natural language processing.In the field of computer vision optical character recognition (OCR) concept was put forward in early 1920s.It is a representative in the field of pattern recognition research important topic.Recently，RBM(Restricted Boltzmann Machines) becomes increasingly popular because of its fast learning algorithm，Contrastive Divergence.For the theoretic perspective，the success of RBM has greatly encouraged the research about the stochastic approximation theory，energybased models and unnormalized statistical models.For the application perspective，RBM has been successfully applied in various machine learning domains，such as classification，regression，dimension reduction，highdimensional time series modeling，sparse overcomplete representations，image transformations and collaborative filtering.In this article，we proposed a new compositional DBMELM network model based on deep Restricted Boltzmann machines which realize data nonlinear dimension reduction and Extreme Learning Machines algorithm.This model can simplify the highdimensional data into lowdimensional space better and then get a better classification effect to represent complicated functions.Finally the experiments on face and handwritten digit recognition prove the superiority.
Liu Shilei，Cui Xiaoming，Nie Ru*.
The DBMELM deep network model[J]. Journal of Nanjing University(Natural Sciences), 2017, 53(3): 497
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
References
[1] 吴晓婷，闫德勤．数据降维方法分析与研究．计算机应用研究，2009，26(8)：2832－2835.(Wu X T，Yan D Q．Analysis and research on method of data dimensionality reduction.Application Research of Computers，2009，26(8)：2832－2835.)
[2] Roweis S T，Saul L K．Nonlinear dimensionality reduction by locally linear embedding.Science，2000，2(9)：2323－2326.
[3] Yang B，Xiang M，Zhang Y．Multimanifold Discriminant Isomap for visualization and classification.Pattern Recognition，2016，55：215－230.
[4] Hettiarachchi R，Peters J F．Multimanifold LLE learning in pattern recognition.Pattern Recognition，2015，48(9)：2947－2960.
[5] Belkinm，Niyogi P．Laplacian eigenmaps for dimensionality reduction and data representation.Neural Computation，2003，15(6)：1373－1396.
[6] Marc’Aurelio R，Christopher P，Sumit C．Efficient learning of sparse representations with an energybased model.In：Platt J．Advances in Neural Information Processing Systems(NIPS 2006)．Cambridge，USA：MIT Press，2006：1137－1144.
[7] Hinton G，Salakhutdinov R．Reducing the dimensionality of data with neural network.Science，2006，6(313)：504－507.
[8] Earnest P I，Krishna M C．Classification of human actions using posebased features and stacked auto encoder.Pattern Recognition Letters，2016，11(83)：268－277.
[9] Liu W，Ma T，Tao D．HSAE：A Hessian regularized sparse autoencoders.Neurocomputing，2016，1(187)：59－65.
[10] Yoshua B，Pascal L，Dan P．Greedy layer wise training of deep networks.In：Platt J．Advances in Neural Information Processing Systems(NIPS 2006)．Cambridge，USA：MIT Press，2006：153－160.
[11] Huang G B，Zhu Q Y，Siew C K．Extreme learning machine：A new learning scheme of feed forward neural networks.In：Dave R N，Sudkamp T．The International Joint Conference on Neural Net Works(IJCNN 2004)．Budapest，Hungary：IEEE Press，2004：985－990.
[12] Huang G B，Zhu Q Y，Siew C K．Extreme learning machine：Theory and applications.Neuro Computing，2006，70(1－3)：489－501.
[13] 胡昭华，宋耀良．基于Autoencoder网络的数据降维和重构．电子与信息学报，2009，31(5)：1189－1192.(Hu Z H，Song Y L．Dimensionality reduction and reconstruction of data based on autoendcoder network.Journal of Electronics & Information Technology，2009，31(5)：1189－1192.)
[14] 刘建伟，刘 媛，罗雄麟．玻尔兹曼机的研究进展．计算机研究与发展.2014，51(1)：1－6.(Liu J W，Liu Y，Luo X L．Research and development on Boltzmann machine.Journal of Computer Research and Development，2014，51(1)：1－6.)
[15] 马 勇，鲍长春，夏丙寅．基于辨别性深度信念网络的说话人分割．清华大学学报(自然科学版)，2013，53(6)：804－807.(Ma Y，Bao C C，Xia B Y．Speaker segmentation base on discriminative deep belief networks.Journal of Tsinghua University(Science & Technology)，2013，53(6)：804－807.)
[16] Hinton G，Osindero S，The Y．A fast learning algorithm for deep belief nets.Neural Computation，2006，18(7)：1527－1554.
PDF(2078265 KB)
