Abstract
Neural networks are the most evolving artificial intelligence method in recent times and have been used for the most complex cognitive tasks. The success of these models has re-scripted many of the benchmark tests in a wide array of fields such as image recognition, natural language processing and speech recognition. The state of the art models leverage on a large amount of labelled training data and a complex model with a huge number of parameters to achieve good results. In this paper, we present a regularization methodology for reducing the size of these complex models while still maintaining generalizability of shallow and deep neural networks. The regularization is based on the monitoring of partial gradients of the loss function with respect to weight parameters. Another way to look at it is the percentage learning evident in a mini-batch training epoch and thereafter removing the weight connections where a certain percentage of learning is not evident. Subsequently, the method is evaluated on several benchmark classification tasks with a drastically smaller size and better performance to models trained with other similar regularization technique of DropConnect.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Bezdek, J.C., Keller, J.M., Krishnapuram, R., Kuncheva, L.I., Pal, N.R.: Will the real iris data please stand up? IEEE Trans. Fuzzy Syst. 7(3), 368–369 (1999). https://doi.org/10.1109/91.771092
Clevert, D.A., Unterthiner, T., Hochreiter, S.: Fast and accurate deep network learning by exponential linear units (elus). arXiv preprint arXiv:1511.07289 (2015)
Glorot, X., Bengio, Y.: Understanding the difficulty of training deep feedforward neural networks. In: Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics, pp. 249–256 (2010)
Goodfellow, I., Bengio, Y., Courville, A.: Deep Learning, vol. 1. MIT Press Cambridge, Cambridge (2016)
Hinton, G.E., Osindero, S., Teh, Y.W.: A fast learning algorithm for deep belief nets. Neural comput. 18(7), 1527–1554 (2006)
Kingma, D.P., Ba, J.L.: Adam: Amethod for stochastic optimization. In: Proceedings of the 3rd International Conference on Learning Representations (ICLR) (2014)
LeCun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521(7553), 436–444 (2015). https://doi.org/10.1038/nature14539
LeCun, Y., Cortes, C., Burges, C.J.: MNIST handwritten digit database. AT&T Labs2 (2010). http://yann.lecun.com/exdb/mnist
Maas, A.L., Hannun, A.Y., Ng, A.Y.: Rectifier nonlinearities improve neural network acoustic models. In: Proceedings of the ICML, vol. 30, p. 3 (2013)
Nair, V., Hinton, G.E.: Rectified linear units improve restricted boltzmann machines. In: Proceedings of the 27th International Conference on Machine Learning (ICML-10), pp. 807–814 (2010)
Srivastava, N., Hinton, G., Krizhevsky, A., Sutskever, I., Salakhutdinov, R.: Dropout: a simple way to prevent neural networks from overfitting. J. Mach. Learn. Res. 15(1), 1929–1958 (2014)
Wan, L., Zeiler, M., Zhang, S., Le Cun, Y., Fergus, R.: Regularization of neural networks using dropconnect. In: International Conference on Machine Learning, pp. 1058–1066 (2013)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Chadha, G.S., Meydani, E., Schwung, A. (2020). Regularizing Neural Networks with Gradient Monitoring. In: Oneto, L., Navarin, N., Sperduti, A., Anguita, D. (eds) Recent Advances in Big Data and Deep Learning. INNSBDDL 2019. Proceedings of the International Neural Networks Society, vol 1. Springer, Cham. https://doi.org/10.1007/978-3-030-16841-4_21
Download citation
DOI: https://doi.org/10.1007/978-3-030-16841-4_21
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-16840-7
Online ISBN: 978-3-030-16841-4
eBook Packages: Intelligent Technologies and RoboticsIntelligent Technologies and Robotics (R0)