Abstract
In this paper, we demonstrate how complex deep learning structures can be understood by humans, if likened to isolated but understandable concepts that use the architecture of Nonnegativity Constrained Autoencoder (NCAE). We show that by constraining most of the weights in the network to be nonnegative using both \(L_1\) and \(L_2\) nonnegativity penalization, a more understandable structure can result with minute deterioration in classification accuracy. Also, this proposed approach yields a more sparse feature extraction and additional output layer sparsification. The concept is illustrated using MNIST and the NORB datasets.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsReferences
Hosseini-Asl, E., Zurada, J., Nasraoui, O.: Deep learning of part-based representation of data using sparse autoencoders with nonnegativity constraints. IEEE Trans. Neural Netw. Learn. Syst. 99, 1–13 (2015)
Bengio, Y., Lamblin, P., Popovici, D., Larochelle, H.: Greedy layer-wise training of deep networks. Adv. Neural Inf. Process. Syst. 19, 153 (2007)
Hinton, G., Salakhutdinov, R.: Reducing the dimensionality of data with neural networks. Science 313(5786), 504–507 (2006)
Vincent, P., Larochelle, H., Bengio, Y., Manzagol, P.: Extracting and composing robust features with denoising autoencoders. In: 25th International Conference on Machine learning, pp. 1096–1103. ACM (2008)
Lee, H., Ekanadham, C., Ng, A.: Sparse deep belief net model for visual area V2. Adv. Neural Inf. Process. Syst. 7, 873–880 (2007)
Lee, D.D., Seung, H.S.: Learning the parts of objects by non-negative matrix factorization. Nature 401(6755), 788–791 (1999)
Chorowski, J., Zurada, J.M.: Learning understandable neural networks with nonnegative weight constraints. IEEE Trans. Neural Netw. Learn. Syst. 26(1), 62–69 (2015)
Olshausen, B.A., et al.: Emergence of simple-cell receptive field properties by learning a sparse code for natural images. Nature 381(6583), 607–609 (1996)
Ranzato, M., Boureau, Y., LeCun, Y.: Sparse feature learning for deep belief networks. Adv. Neural Inf. Process. Syst. 20, 1185–1192 (2007)
Ishikawa, M.: Structural learning with forgetting. Neural Netw. 9(3), 509–521 (1996)
Bartlett, P.L.: The sample complexity of pattern classification with neural networks: the size of the weights is more important than the size of the network. IEEE Trans. Inf. Theory 44(2), 525–536 (1998)
Ayinde, B.O., Barnawi, A.Y.: Differential evolution based deployment of wireless sensor networks. In: 2014 IEEE/ACS 11th International Conference on Computer Systems and Applications (AICCSA), pp. 131–137. IEEE (2014)
Gnecco, G., Sanguineti, M.: Regularization techniques and suboptimal solutions to optimization problems in learning from data. Neural Comput. 22(3), 793–829 (2010)
Moody, J., Hanson, S., Krogh, A., Hertz, J.A.: A simple weight decay can improve generalization. Adv. Neural Inf. Process. Syst. 4, 950–957 (1995)
Lemme, A., Reinhart, R., Steil, J.: Online learning and generalization of parts-based image representations by non-negative sparse autoencoders. Neural Netw. 33, 194–203 (2012)
Nguyen, T.D., Tran, T., Phung, D., Venkatesh, S.: Learning partsbased representations with nonnegative restricted Boltzmann machine. In: Asian Conference on Machine Learning, pp. 133–148 (2013)
Wright, S.J., Nocedal, J.: Numerical Optimization. Springer, New York (1999)
Hashim, H.A., Ayinde, B., Abido, M.: Optimal placement of relay nodes in wireless sensor network using artificial bee colony algorithm. J. Netw. Comput. Appl. 64, 239–248 (2016)
Zurada, J.M.: Introduction to Artificial Neural Systems. West Publishing Co., St. Paul (1992)
Hinton, G., Osindero, S., Teh, Y.W.: A fast learning algorithm for deep belief nets. Neural Comput. 18(7), 1527–1554 (2006)
LeCun, Y., Cortes, C., Burges, C.J.: The MNIST database of handwritten digits (1998)
der Maaten, L.V., Hinton, G.: Visualizing data using t-SNE. J. Mach. Learn. Res. 9(11), 2579–2605 (2008)
Hinton, G.E., Srivastava, N., Krizhevsky, A., Sutskever, I., Salakhutdinov, R.R.: Improving neural networks by preventing co-adaptation of feature detectors (2012). arXiv preprint arXiv:1207.0580
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer International Publishing Switzerland
About this paper
Cite this paper
Ayinde, B.O., Hosseini-Asl, E., Zurada, J.M. (2016). Visualizing and Understanding Nonnegativity Constrained Sparse Autoencoder in Deep Learning. In: Rutkowski, L., Korytkowski, M., Scherer, R., Tadeusiewicz, R., Zadeh, L., Zurada, J. (eds) Artificial Intelligence and Soft Computing. ICAISC 2016. Lecture Notes in Computer Science(), vol 9692. Springer, Cham. https://doi.org/10.1007/978-3-319-39378-0_1
Download citation
DOI: https://doi.org/10.1007/978-3-319-39378-0_1
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-39377-3
Online ISBN: 978-3-319-39378-0
eBook Packages: Computer ScienceComputer Science (R0)