Abstract
A new procedure is proposed to simplify a Shallow Neural Network. With such aim, we introduce the concept of “path weights”, which consists of the multiplication of the successive weights along a single path from the input layer to the output layer. This concept is used alongside a direct analysis of the calculations performed by a neural network to simplify a neural network via removing paths determined to be not relevant for decision making.
This study compares the proposed “path weights” method for network lightening with other methods, taken from input ranking, namely the Garson, and Paliwal and Kumar.
All the different methods of network lightening reduce the network complexity, favoring interpretability and keeping the prediction at high levels, well above randomness. However, Garson’s and Paliwal and Kumar’s methods maintain the totality of the weights between the hidden layer and the output layer, which keeps the number of connections high and reduces the capability of analyzing hidden node importance in the network’s decision-making. Yet, the proposed method based on “path weights” provides lighter networks while retaining a high prediction rate, thus providing a high hit per network connection ratio. Moreover, the lightened neural network inputs are coherent with the established literature, identifying the brain regions that participate in motor execution.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Haynes, J.-D., Rees, G.: Decoding mental states from brain activity in humans. Nat. Rev. Neurosci. 7, 523–534 (2006). https://doi.org/10.1038/nrn1931
Hanson, S.J., Matsuka, T., Haxby, J.V.: Combinatorial codes in ventral temporal lobe for object recognition: Haxby (2001) revisited: is there a “face” area? Neuroimage 23, 156–166 (2004). https://doi.org/10.1016/j.neuroimage.2004.05.020
Onut, I.-V., Ghorbani, A.A.: Classifying cognitive states from fMRI data using neural networks. In: Proceedings. 2004 IEEE International Joint Conference on Neural Networks, pp. 2871–2875 (2004). https://doi.org/10.1109/IJCNN.2004.1381114
Sona, D., Veeramachaneni, S., Olivetti, E., Avesani, P.: Inferring cognition from fMRI brain images. In: de Sá, J.M., Alexandre, L.A., Duch, W., Mandic, D. (eds.) ICANN 2007. LNCS, vol. 4669, pp. 869–878. Springer, Heidelberg (2007). https://doi.org/10.1007/978-3-540-74695-9_89
Marques dos Santos, J.P., Moutinho, L., Castelo-Branco, M.: ‘Mind reading’: hitting cognition by using ANNs to analyze fMRI data in a paradigm exempted from motor responses. In: International Workshop on Artificial Neural Networks and Intelligent Information Processing (ANNIIP 2014), pp. 45–52. Scitepress (Science and Technology Publications, Lda.), Vienna, Austria (2014). https://doi.org/10.5220/0005126400450052
Weygandt, M., Stark, R., Blecker, C., Walter, B., Vaitl, D.: Real-time fMRI pattern-classification using artificial neural networks. Clin. Neurophysiol. 118, e114–e114 (2007). https://doi.org/10.1016/j.clinph.2006.11.265
de Oña, J., Garrido, C.: Extracting the contribution of independent variables in neural network models: a new approach to handle instability. Neural Comput. Appl. 25(3–4), 859–869 (2014). https://doi.org/10.1007/s00521-014-1573-5
Adadi, A., Berrada, M.: Peeking inside the black-box: a survey on explainable artificial intelligence (XAI). IEEE Access 6, 52138–52160 (2018). https://doi.org/10.1109/ACCESS.2018.2870052
Samek, W., Müller, K.-R.: Towards explainable artificial intelligence. In: Samek, W., Montavon, G., Vedaldi, A., Hansen, L.K., Müller, K.-R. (eds.) Explainable AI: Interpreting, Explaining and Visualizing Deep Learning. LNCS (LNAI), vol. 11700, pp. 5–22. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-28954-6_1
Tjoa, E., Guan, C.: A survey on explainable artificial intelligence (XAI): toward medical XAI. IEEE Trans. Neural Netw. Learn. Syst. 32, 4793–4813 (2021). https://doi.org/10.1109/TNNLS.2020.3027314
Blalock, D., Gonzalez Ortiz, J.J., Frankle, J., Guttag, J.: What is the state of neural network pruning? In: Dhillon, I., Papailiopoulos, D., Sze, V. (eds.) 3rd Conference on Machine Learning and Systems, MLSys 2020, vol. 2, pp. 129-146, Austin (TX), USA (2020)
Zhao, F., Zeng, Y.: Dynamically optimizing network structure based on synaptic pruning in the brain. Frontiers in Systems Neuroscience 15, 620558 (2021). https://doi.org/10.3389/fnsys.2021.620558
Mirkes, E.M.: Artificial neural network pruning to extract knowledge. In: 2020 International Joint Conference on Neural Networks (IJCNN), pp. 1–8. IEEE (2020). https://doi.org/10.1109/IJCNN48605.2020.9206861
Olden, J.D., Joy, M.K., Death, R.G.: An accurate comparison of methods for quantifying variable importance in artificial neural networks using simulated data. Ecol. Model. 178, 389–397 (2004). https://doi.org/10.1016/j.ecolmodel.2004.03.013
Garson, D.G.: Interpreting neural network connection weights. AI Expert 6, 46–51 (1991)
Paliwal, M., Kumar, U.A.: Assessing the contribution of variables in feed forward neural network. Appl. Soft Comput. 11, 3690–3696 (2011). https://doi.org/10.1016/j.asoc.2011.01.040
Fischer, A.: How to determine the unique contributions of input-variables to the nonlinear regression function of a multilayer perceptron. Ecol. Model. 309–310, 60–63 (2015). https://doi.org/10.1016/j.ecolmodel.2015.04.015
de Sá, C.R.: Variance-based feature importance in neural networks. In: Kralj Novak, P., Šmuc, T., Džeroski, S. (eds.) Discovery Science. Lecture Notes in Computer Science (Lecture Notes in Artificial Intelligence), vol. 11828, pp. 306–315. Springer, Cham (2019). https://doi.org/10.1007/978-3-030-33778-0_24
Luíza da Costa, N., Dias de Lima, M., Barbosa, R.: Evaluation of feature selection methods based on artificial neural network weights. Expert Syst. Appl. 168, 114312 (2021). https://doi.org/10.1016/j.eswa.2020.114312
Bondarenko, A., Borisov, A., Alekseeva, L.: Neurons vs weights pruning in artificial neural networks. In: 10th International Scientific and Practical Conference on Environment. Technologies. Resources, vol. 3, pp. 22–28. Rēzekne Academy of Technologies, Rēzekne (2015)
Karnin, E.D.: A simple procedure for pruning back-propagation trained neural networks. IEEE Trans. Neural Netw. 1, 239–242 (1990). https://doi.org/10.1109/72.80236
Penfield, W., Boldrey, E.: Somatic motor and sensory representation in the cerbral cortex of man as studied by electrical stimulation. Brain 60, 389–443 (1937). https://doi.org/10.1093/brain/60.4.389
Yeo, B.T.T., et al.: The organization of the human cerebral cortex estimated by intrinsic functional connectivity. J. Neurophysiol. 106, 1125–1165 (2011). https://doi.org/10.1152/jn.00338.2011
Van Essen, D.C., Glasser, M.F.: The human connectome project: progress and prospects. In: Cerebrum 2016, cer-10–16 (2016)
Van Essen, D.C., Smith, S.M., Barch, D.M., Behrens, T.E.J., Yacoub, E., Ugurbil, K.: The WU-Minn human connectome project: an overview. Neuroimage 80, 62–79 (2013). https://doi.org/10.1016/j.neuroimage.2013.05.041
Elam, J.S., et al.: The human connectome project: a retrospective. Neuroimage 244, 118543 (2021). https://doi.org/10.1016/j.neuroimage.2021.118543
Buckner, R.L., Krienen, F.M., Castellanos, A., Diaz, J.C., Yeo, B.T.T.: The organization of the human cerebellum estimated by intrinsic functional connectivity. J. Neurophysiol. 106, 2322–2345 (2011). https://doi.org/10.1152/jn.00339.2011
Beckmann, C.F., Smith, S.M.: Probabilistic independent component analysis for functional magnetic resonance imaging. IEEE Trans. Med. Imaging 23, 137–152 (2004). https://doi.org/10.1109/TMI.2003.822821
Minka, T.P.: Automatic choice of dimensionality for PCA. Technical Report 514, MIT Media Lab Vision and Modeling Group. MIT (2000)
Hyvärinen, A.: Fast and robust fixed-point algorithms for independent component analysis. IEEE Trans. Neural Netw. 10, 626–634 (1999). https://doi.org/10.1109/72.761722
Buckner, R.L.: Event-related fMRI and the hemodynamic response. Hum. Brain Mapp. 6, 373–377 (1998). https://doi.org/10.1002/(SICI)1097-0193(1998)6:5/6%3c373::AID-HBM8%3e3.0.CO;2-P
Limas, M.C., Meré, J.B.O., Marcos, A.G., de Pisón Ascacibar, F., Espinoza, A.P., Elías, F.A.: A MORE flexible neural network package (0.2–12). León (2010)
R Development Core Team: R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna (2010)
Haykin, S.: Neural Networks and Learning Machines. Prentice Hall, New Jersey (2009)
Le Cun, Y.: Efficient learning and second-order methods. Tutorial NIPS 93, 61 (1993)
Acknowledgements
This work was partially financially supported by Base Funding - UIDB/00027/2020 of the Artificial Intelligence and Computer Science Laboratory – LIACC - funded by national funds through the FCT/MCTES (PIDDAC).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
1 Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
Copyright information
© 2023 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Marques dos Santos, J.D., Marques dos Santos, J.P. (2023). Path Weights Analyses in a Shallow Neural Network to Reach Explainable Artificial Intelligence (XAI) of fMRI Data. In: Nicosia, G., et al. Machine Learning, Optimization, and Data Science. LOD 2022. Lecture Notes in Computer Science, vol 13811. Springer, Cham. https://doi.org/10.1007/978-3-031-25891-6_31
Download citation
DOI: https://doi.org/10.1007/978-3-031-25891-6_31
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-25890-9
Online ISBN: 978-3-031-25891-6
eBook Packages: Computer ScienceComputer Science (R0)