Abstract
In machine learning, one of the most efficient feature extraction methods is autoencoder which transforms the data from its original space to a latent space. The transformed data is then used for machine learning downstream tasks rather than the original data. However, there is little research about choosing the best number of latent space dimensions (k) for autoencoders that can affect the result of these tasks. In this paper, we focus on the impact of k on the accuracy of a downstream task. Concretely, we survey recently developed autoencoders and their characteristics, and conduct experiments using different autoencoders and k for extracting information from different datasets. We then present the accuracy of a classifier on the extracted datasets and the reconstruction error of the autoencoders according to k. From the empirical results, we recommend the best k of the latent space dimension for each dataset and each autoencoder.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
The source code is available at: https://github.com/KienMN/Autoencoder-Experiments.
References
Doersch, C.: Tutorial on variational autoencoders. arXiv preprint arXiv:1606.05908 (2016)
Ghojogh, B., et al.: Feature selection and feature extraction in pattern analysis: a literature review. arXiv preprint arXiv:1905.02845 (2019)
Goodfellow, I., Bengio, Y., Courville, A.: Deep Learning. MIT Press (2016). http://www.deeplearningbook.org
Gretton, A., Borgwardt, K., Rasch, M., Schölkopf, B., Smola, A.J.: A kernel method for the two-sample-problem. In: Schölkopf, B., Platt, J.C., Hoffman, T. (eds.) Advances in Neural Information Processing Systems 19, pp. 513–520. MIT Press (2007). http://papers.nips.cc/paper/3110-a-kernel-method-for-the-two-sample-problem.pdf
Kingma, D.P., Welling, M.: Auto-encoding variational Bayes. arXiv preprint arXiv:1312.6114 (2013)
Krizhevsky, A.: Learning multiple layers of features from tiny images. Technical report (2009)
LeCun, Y., Cortes, C., Burges, C.: MNIST handwritten digit database. ATT Labs 2 (2010). http://yann.lecun.com/exdb/mnist
Makhzani, A., Shlens, J., Jaitly, N., Goodfellow, I., Frey, B.: Adversarial autoencoders. arXiv preprint arXiv:1511.05644 (2015)
Tschannen, M., Bachem, O., Lucic, M.: Recent advances in autoencoder-based representation learning. arXiv preprint arXiv:1812.05069 (2018)
Vincent, P., Larochelle, H., Bengio, Y., Manzagol, P.A.: Extracting and composing robust features with denoising autoencoders. In: Proceedings of the 25th International Conference on Machine Learning, pp. 1096–1103 (2008)
Xiao, H., Rasul, K., Vollgraf, R.: Fashion-MNIST: a novel image dataset for benchmarking machine learning algorithms. CoRR abs/1708.07747 (2017). http://arxiv.org/abs/1708.07747
Zhao, S., Song, J., Ermon, S.: InfoVAE: information maximizing variational autoencoders. arXiv preprint arXiv:1706.02262 (2017)
Acknowledgement
This research was supported by Korea Institute of Science and Technology Information (KISTI).
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
Mai Ngoc, K., Hwang, M. (2020). Finding the Best k for the Dimension of the Latent Space in Autoencoders. In: Nguyen, N.T., Hoang, B.H., Huynh, C.P., Hwang, D., Trawiński, B., Vossen, G. (eds) Computational Collective Intelligence. ICCCI 2020. Lecture Notes in Computer Science(), vol 12496. Springer, Cham. https://doi.org/10.1007/978-3-030-63007-2_35
Download citation
DOI: https://doi.org/10.1007/978-3-030-63007-2_35
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-63006-5
Online ISBN: 978-3-030-63007-2
eBook Packages: Computer ScienceComputer Science (R0)