Abstract
The striking results of deep neural networks (DNN) have motivated its wide acceptance to tackle large datasets and complex tasks such as natural language processing, facial recognition, and artificial image generation. However, DNN parameters are often empirically selected on a trial-and-error approach without detailed information on convergence behavior. While some visualization techniques have been proposed to aid the comprehension of general-purpose neural networks, only a few explore the training process, lacking the ability to adequately display how abstract representations are formed and represent the influence of training parameters during this process. This paper describes neural network training fingerprint (NNTF), a visual analytics approach to investigate the training process of any neural network performing classification. NNTF allows understanding how classification decisions change along the training process, displaying information about convergence, oscillations, and training rates. We show its usefulness through case studies and demonstrate how it can support the analysis of training parameters.
Graphical abstract
Similar content being viewed by others
Notes
Source code available at https://github.com/marthadais/NeuralNetworkTrainingFingerprint.
References
Abadi M, Barham P, Chen J, Chen Z, Davis A, Dean J, Devin M, Ghemawat S, Irving G, Isard M, et al. (2016) Tensorflow: a system for large-scale machine learning. In: OSDI, vol 16, pp 265–283
Alencar AB, Börner K, Paulovich FV, de Oliveira MCF (2012) Time-aware visualization of document collections. In: Proceedings of the 27th annual ACM symposium on applied computing, SAC ’12. Association for Computing Machinery, New York, pp 997–1004. https://doi.org/10.1145/2245276.2245469
Babiker HKB, Goebel R (2017) An introduction to deep visual explanation. ArXiv preprint arXiv:1711.09482
Bach S, Binder A, Montavon G, Klauschen F, Müller KR, Samek W (2015) On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance propagation. PLoS One 10(7):e0130140
Borg I, Groenen P (2003) Modern multidimensional scaling: theory and applications. J Educ Meas 40(3):277–280
Cantareira GD, Etemad E, Paulovich FV (2020) Exploring neural network hidden layer activity using vector fields. Information 11(9):426
Chen C, Yuan J, Lu Y, Liu Y, Su H, Yuan S, Liu S (2020) Oodanalyzer: interactive analysis of out-of-distribution samples. IEEE Trans Vis Comput Graph 27(7):3335–3349
de Araújo Tiburtino Neves TT, Martins RM, Coimbra DB, Kucher K, Kerren A, Paulovich FV (2021) Fast and reliable incremental dimensionality reduction for streaming data. Comput Graph. https://doi.org/10.1016/j.cag.2021.08.009
Erhan D, Bengio Y, Courville A, Vincent P (2009) Visualizing higher-layer features of a deep network. Univ Montr 1341(3):1
Goodfellow I, Bengio Y, Courville A, Bengio Y (2016) Deep learning. MIT Press, Cambridge
Gschwandtner T, Erhart O (2018) Know your enemy: identifying quality problems of time series data. In: 2018 IEEE Pacific visualization symposium (PacificVis). IEEE, pp 205–214
Gu D, Li Y, Jiang F, Wen Z, Liu S, Shi W, Lu G, Zhou C (2020) Vinet: a visually interpretable image diagnosis network. IEEE Trans Multimedia 22(7):1720–1729
Hohman F, Kahng M, Pienta R, Chau DH (2018) Visual analytics in deep learning: an interrogative survey for the next frontiers. IEEE Trans Vis Comput Graph 25(8):2674–2693
Hohman F, Wongsuphasawat K, Kery MB, Patel K (2020) Understanding and visualizing data iteration in machine learning. In: Proceedings of the 2020 CHI conference on human factors in computing systems, pp 1–13
Kahng M, Andrews PY, Kalro A, Chau DHP (2018) Activis: visual exploration of industry-scale deep neural network models. IEEE Trans Vis Comput Graph 24(1):88–97
Krizhevsky A (2010) Convolutional deep belief networks on cifar-10. Unpublished manuscript 40(7):1–9
Krizhevsky A, Sutskever I, Hinton GE (2012) Imagenet classification with deep convolutional neural networks. In: Advances in neural information processing systems, pp 1097–1105
Krogh A, Hertz JA (1992) A simple weight decay can improve generalization. In: Advances in neural information processing systems, pp 950–957
LeCun Y, Bengio Y, Hinton G (2015) Deep learning. Nature 521(7553):436
LeCun Y, Haffner P, Bottou L, Bengio Y (1999) Object recognition with gradient-based learning. In: Shape, contour and grouping in computer vision. Springer, Berlin, Heidelberg, pp 319–345. https://doi.org/10.1007/3-540-46805-6_19
Liu D, Cui W, Jin K, Guo Y, Qu H (2018) Deeptracker: visualizing the training process of convolutional neural networks. ACM Trans Intell Syst Technol (TIST) 10(1):1–25
Liu M, Shi J, Cao K, Zhu J, Liu S (2017) Analyzing the training processes of deep generative models. IEEE Trans Vis Comput Graph 24(1):77–87
Liu S, Wang X, Liu M, Zhu J (2017) Towards better analysis of machine learning models: a visual analytics perspective. Vis Inf 1(1):48–56
Ma Y, Xie T, Li J, Maciejewski R (2019) Explaining vulnerabilities to adversarial machine learning through visual analytics. IEEE Trans Vis Comput Graph 26(1):1075–1085
Mahendran A, Vedaldi A (2016) Visualizing deep convolutional neural networks using natural pre-images. Int J Comput Vis 120(3):233–255
Pezzotti N, Höllt T, Van Gemert J, Lelieveldt BP, Eisemann E, Vilanova A (2018) Deepeyes: progressive visual analytics for designing deep neural networks. IEEE Trans Vis Comput Graph 24(1):98–108
Rauber PE, Fadel SG, Falcao AX, Telea AC (2017) Visualizing the hidden activity of artificial neural networks. IEEE Trans Vis Comput Graph 23(1):101–110
Samek W, Binder A, Montavon G, Lapuschkin S, Müller KR (2017) Evaluating the visualization of what a deep neural network has learned. IEEE Trans Neural Netw Learn Syst 28(11):2660–2673
Scherer D, Müller A, Behnke S (2010) Evaluation of pooling operations in convolutional architectures for object recognition. In: International conference on artificial neural networks, pp 92–101. Springer
Shang W, Sohn K, Almeida D, Lee H (2016) Understanding and improving convolutional neural networks via concatenated rectified linear units. In: International conference on machine learning, pp 2217–2225
Simonyan K, Vedaldi A, Zisserman A (2014) Deep inside convolutional networks: visualising image classification models and saliency maps. In: 2nd international conference on learning representations (ICLR), workshop track proceedings
Simonyan K, Zisserman A (2014) Very deep convolutional networks for large-scale image recognition. ArXiv preprint arXiv:1409.1556
Srivastava N, Hinton G, Krizhevsky A, Sutskever I, Salakhutdinov R (2014) Dropout: a simple way to prevent neural networks from overfitting. J Mach Learn Res 15(1):1929–1958
Takens F et al (1981) Detecting strange attractors in turbulence. Lect Notes Math 898(1):366–381
Van Der Maaten L (2014) Accelerating t-SNE using tree-based algorithms. J Mach Learn Res 15(1):3221–3245
Wang J, Gou L, Shen HW, Yang H (2018) Dqnviz: a visual analytics approach to understand deep q-networks. IEEE Trans Vis Comput Graph 25(1):288–298
Webber CL Jr, Zbilut JP (1994) Dynamical assessment of physiological systems and states using recurrence plot strategies. J Appl Physiol 76(2):965–973
Yosinski J, Clune J, Nguyen A, Fuchs T, Lipson H (2015) Understanding neural networks through deep visualization. arXiv preprint arXiv:1506.06579
Zeiler MD, Fergus R (2014) Visualizing and understanding convolutional networks. In: European conference on computer vision, pp 818–833. Springer
Zintgraf LM, Cohen TS, Adel T, Welling M (2017) Visualizing deep neural network decisions: prediction difference analysis. ArXiv preprint arXiv:1702.04595
Acknowledgements
We acknowledge the support of the Natural Sciences and Engineering Research Council of Canada (NSERC).
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Ferreira, M.D., Cantareira, G.D., de Mello, R.F. et al. Neural network training fingerprint: visual analytics of the training process in classification neural networks. J Vis 25, 593–612 (2022). https://doi.org/10.1007/s12650-021-00809-4
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12650-021-00809-4