Learning parts-based representations with nonnegative restricted boltzmann machine

Nguyen, Tu Dinh, Tran, Truyen, Phung, Dinh and Venkatesh, Svetha 2013, Learning parts-based representations with nonnegative restricted boltzmann machine, in ACML 2013 : Proceedings of the 5th Asian Conference on Machine Learning, [The Conference], [Canberra, A.C.T.], pp. 133-148.

Attached Files
Name Description MIMEType Size Downloads

Title Learning parts-based representations with nonnegative restricted boltzmann machine
Author(s) Nguyen, Tu Dinh
Tran, TruyenORCID iD for Tran, Truyen orcid.org/0000-0001-6531-8907
Phung, DinhORCID iD for Phung, Dinh orcid.org/0000-0002-9977-8247
Venkatesh, SvethaORCID iD for Venkatesh, Svetha orcid.org/0000-0001-8675-6631
Conference name Asian Conference on Machine Learning (5th : 2013 : Canberra, Australian Capital Territory))
Conference location Canberra, Australian Capital Territory
Conference dates 2013/11/13 - 2013/11/15
Title of proceedings ACML 2013 : Proceedings of the 5th Asian Conference on Machine Learning
Publication date 2013
Start page 133
End page 148
Total pages 16
Publisher [The Conference]
Place of publication [Canberra, A.C.T.]
Keyword(s) parts-based representation
restricted Boltzmann machines
learning representation
semantic features
Summary The success of any machine learning system depends critically on effective representations of data. In many cases, especially those in vision, it is desirable that a representation scheme uncovers the parts-based, additive nature of the data. Of current representation learning schemes, restricted Boltzmann machines (RBMs) have proved to be highly effective in unsupervised settings. However, when it comes to parts-based discovery, RBMs do not usually produce satisfactory results. We enhance such capacity of RBMs by introducing nonnegativity into the model weights, resulting in a variant called nonnegative restricted Boltzmann machine (NRBM). The NRBM produces not only controllable decomposition of data into interpretable parts but also offers a way to estimate the intrinsic nonlineardimensionality of data. We demonstrate the capacity of our model on well-known datasets of handwritten digits, faces and documents. The decomposition quality on images is comparable with or better than what produced by the nonnegative matrix factorisation (NMF), and the thematic features uncovered from text are qualitatively interpretable in a similar manner to that of the latent Dirichlet allocation (LDA). However, the learnt features, when used for classification, are more discriminative than those discovered by both NMF and LDA and comparable with those by RBM.
Language eng
Field of Research 080109 Pattern Recognition and Data Mining
080702 Health Informatics
Socio Economic Objective 0 Not Applicable
HERDC Research category E1.1 Full written paper - refereed
Copyright notice ©2013, The Authors
Persistent URL http://hdl.handle.net/10536/DRO/DU:30081487

Connect to link resolver
Unless expressly stated otherwise, the copyright for items in DRO is owned by the author, with all rights reserved.

Version Filter Type
Citation counts: TR Web of Science Citation Count  Cited 0 times in TR Web of Science
Scopus Citation Count Cited 8 times in Scopus
Google Scholar Search Google Scholar
Access Statistics: 574 Abstract Views, 2 File Downloads  -  Detailed Statistics
Created: Fri, 19 Feb 2016, 11:56:59 EST

Every reasonable effort has been made to ensure that permission has been obtained for items included in DRO. If you believe that your rights have been infringed by this repository, please contact drosupport@deakin.edu.au.