Abstract
The use of deep networks has improved the state of the art in various domains of AI, making practical applications possible. At the same time, there are increasing calls to make learning systems more transparent and explainable, due to concerns that they might develop biases in their internal representations that might lead to unintended discrimination, when applied to sensitive personal decisions. The use of vast subsymbolic distributed representations has made this task very difficult. We suggest that we can learn a lot about the biases and the internal representations of a deep network without having to unravel its connections, but by adopting the old psychological approach of analysing its “slips of the tongue”. We demonstrate in a practical example that an analysis of the confusion matrix can reveal that a CNN has represented a biological task in a way that reflects our understanding of taxonomy, inferring more structure than it was requested to by the training algorithm. In particular, we show how a CNN trained to recognise animal families, contains also higher order information about taxa such as the superfamily, parvorder, suborder and order for example. We speculate that various forms of psycho-metric testing for neural networks might provide us insight about their inner workings.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
The ResNet CNN used in their paper had the same architecture, but was much deeper. We trained a similar 152-layer network to [9] but found no clear difference with our 18-layer model, which aimed to strike a better balance between the depth of the network and the associated computational load.
References
Alain, G., Bengio, Y.: Understanding intermediate layers using linear classifier probes. arXiv preprint arXiv:1610.01644 (2016)
Chen, G., Han, T.X., He, Z., Kays, R., Forrester, T.: Deep convolutional neural network based species recognition for wild animal monitoring. In 2014 IEEE International Conference on Image Processing (ICIP), pp. 858–862. IEEE (2014)
Datta, A., Sen, S., Zick, Y.: Algorithmic transparency via quantitative input influence: theory and experiments with learning systems. In: 2016 IEEE Symposium on Security and Privacy (SP), pp. 598–617. IEEE (2016)
Day, W.H.E., Edelsbrunner, H.: Efficient algorithms for agglomerative hierarchical clustering methods. J. Classif. 1(1), 7–24 (1984)
Denker, J., Schwartz, D., Wittner, B., Solla, S., Howard, R., Jackel, L., Hopfield, J.: Large automatic learning, rule extraction, and generalization. Complex Syst. 1(5), 877–922 (1987)
Federhen, S.: The NCBI taxonomy database. Nucleic Acids Res. 40(D1), D136–D143 (2012)
Felsenstein, J.: Inferring Phylogenies, vol. 2
Freitas, A.A.: Comprehensible classification models: a position paper. ACM SIGKDD Explor. Newsl. 15(1), 1–10 (2014)
Gómez, A., Salazar, A., Vargas, F.: Towards automatic wild animal monitoring: identification of animal species in camera-trap images using very deep convolutional neural networks. arXiv preprint arXiv:1603.06169 (2016)
Goodman, B., Flaxman, S.: European union regulations on algorithmic decision-making and a “right to explanation”. arXiv preprint arXiv:1606.08813 (2016)
Gunning, D.: Explainable artificial intelligence (XAI). Defense Advanced Research Projects Agency (DARPA) (2017)
Hassibi, B., Stork, D.G., Wolff, G.J.: Optimal brain surgeon and general network pruning. In: IEEE International Conference on Neural Networks, vol. 1, pp. 293–299 (1993)
He, K., Zhang, X., Ren, S., Sun, J.: Deep residual learning for image recognition. CoRR, abs/1512.03385 (2015)
Jia, S., Lansdall-Welfare, T., Cristianini, N.: Gender classification by deep learning on millions of weakly labelled images. In: 2016 IEEE 16th International Conference on Data Mining Workshops (ICDMW), pp. 462–467. IEEE (2016)
Jung, H., Lee, S., Park, S., Lee, I., Ahn, C., Kim, J.: Deep temporal appearance-geometry network for facial expression recognition. arXiv preprint arXiv:1503.01532 (2015)
Krause, J., Perer, A., Bertini, E.: Using visual analytics to interpret predictive machine learning models. arXiv preprint arXiv:1606.05685 (2016)
Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep convolutional neural networks. In: Advances in Neural Information Processing Systems, pp. 1097–1105 (2012)
Le, Q.V.: Building high-level features using large scale unsupervised learning. In: 2013 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 8595–8598. IEEE (2013)
Lecun, Y., Bengio, Y., Hinton, G.: Deep learning. Nature 521(7553), 436–444 (2015)
MathWorks. Bioinformatics toolbox: User’s guide (r2017a) (2017). www.mathworks.com/help/pdf_doc/bioinfo/bioinfo_ug.pdf
Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A.A., Veness, J., Bellemare, M.G., Graves, A., Riedmiller, M., Fidjeland, A.K., Ostrovski, G., et al.: Human-level control through deep reinforcement learning. Nature 518(7540), 529–533 (2015)
Rissanen, J.: Minimum description length principle. Wiley Online Library (1985)
Silver, D., Huang, A., Maddison, C.J., Guez, A., Sifre, L., Van Den Driessche, G., Schrittwieser, J., Antonoglou, I., Panneershelvam, V., Lanctot, M., et al.: Mastering the game of go with deep neural networks and tree search. Nature 529(7587), 484–489 (2016)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2017 Springer International Publishing AG
About this paper
Cite this paper
Jia, S., Lansdall-Welfare, T., Cristianini, N. (2017). Freudian Slips: Analysing the Internal Representations of a Neural Network from Its Mistakes. In: Adams, N., Tucker, A., Weston, D. (eds) Advances in Intelligent Data Analysis XVI. IDA 2017. Lecture Notes in Computer Science(), vol 10584. Springer, Cham. https://doi.org/10.1007/978-3-319-68765-0_12
Download citation
DOI: https://doi.org/10.1007/978-3-319-68765-0_12
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-68764-3
Online ISBN: 978-3-319-68765-0
eBook Packages: Computer ScienceComputer Science (R0)