近年來，隨著人工智慧的迅速發展，人工智慧改變了我們的生活和許多領域，其造成的影響難以量化。在一些領域中的表現甚至已經超越了人類，例如圍棋、象棋、德州撲克等遊戲，但經常浮現出來的問題是，人工智慧的決策過程往往是黑箱，那麼它又是如何做出決定的呢？
本研究提出了一種基於卷積神經網路的深度學習模型，利用大腦視覺皮質的運作方式和階層架構及時序性的概念，來解釋深度學習模型的決策過程。此模型使用多層的架構進行影像分類，當影像輸入後，透過高斯卷積及特徵增強的機制，並將影像特徵透過時序性進行結合並輸出至下一層，就如同如視覺皮質在接收影像訊號時的運作方式，底層神經元會將細小的資訊根據時序性進行結合，並透過階層性的結構進行傳遞，最後使用一層全連接層，將其輸出轉換為影像的分類結果。
本實驗中，共使用兩種資料集，分別是 MNIST 和 Fashion-MNIST，皆有不錯的表現。在每一階段針對特徵進行解釋，並透過特徵視覺化，可以觀察到每一層的特徵都有獨特的意義，這對於可解釋的人工智慧具有重要意義，同時為機器學習和相關領域的發展提供了新的思路和方法。

In recent years, with the rapid development of artificial intelligence (AI), it has significantly transformed our lives and various domains, and its impact is difficult to quantify. AI has even surpassed humans in performance in certain areas such as Go, chess, and Texas Hold’em poker. However, the decision- making process of artificial intelligence (AI) is often considered a black box, raising the question of how it actually makes decisions.
This research proposes a deep learning model based on convolutional neural networks (CNNs) that incorporates the concepts of multi-layer SOM and the functioning of the visual cortex in the human brain to provide interpretability to the decision-making process of deep learning models. This model uses a multi-layer architecture for image classification. When an image is inputted, it undergoes Gaussian convolution and feature enhancement mechanisms. The image features are then combined in a temporal sequence and propagated to the next layer, mimicking the operation of the visual cortex in processing visual signals. Lower-level neurons integrate fine-grained information and transmit it hierarchically through the network structure. Finally, a fully connected layer is used to convert the output into the classification result of the image.
In our experiment, two datasets, namely MNIST and Fashion-MNIST, were used, both yielding favorable performance. At each stage, the features were explained, and through feature visualization, it was observed that each layer had its unique significance. This is of paramount importance for explainable AI, providing new insights and methods for the development of machine learning and related fields.

[1] E. R. Kandel, J. H. Schwartz, T. M. Jessell, S. Siegelbaum, A. J. Hudspeth, S. Mack, et al., Principles of neural science. McGraw-hill New York, 2000, vol. 4.
[2] S. Herculano-Houzel, “The human brain in numbers: A linearly scaled-up primate brain,”Frontiers in human neuroscience, p. 31, 2009.
[3] B. Fischl and A. M. Dale, “Measuring the thickness of the human cerebral cortex from magnetic resonance images,” Proceedings of the National Academy of Sciences, vol. 97, no. 20, pp. 11 050–11 055, 2000.
[4] J. Hawkins and S. Blakeslee, On intelligence, trans. by 洪蘭. Macmillan, 2004.
[5] Y. LeCun, L. Bottou, Y. Bengio, and P. Haffner, “Gradient-based learning applied to document recognition,” Proceedings of the IEEE, vol. 86, no. 11, pp. 2278–2324, 1998.
[6] T. Kohonen, “The self-organizing map,” Proceedings of the IEEE, vol. 78, no. 9, pp. 1464–1480, 1990.
[7] Y.-Y. Hsu et al., “基於多層自我組織映射圖之可視覺化深度學習模型,” Master’s thesis, National Central University, 2018.
[8] D. Gunning and D. Aha, “Darpa＇s explainable artificial intelligence (xai) program,” AI magazine, vol. 40, no. 2, pp. 44–58, 2019.
[9] P. Jackson, “Introduction to expert systems,” 1986.
[10] A. B. Arrieta, N. Díaz-Rodríguez, J. Del Ser, et al., “Explainable artificial intelligence(xai): Concepts, taxonomies, opportunities and challenges toward responsible ai,” Information fusion, vol. 58, pp. 82–115, 2020.
[11] I. E. Nielsen, D. Dera, G. Rasool, R. P. Ramachandran, and N. C. Bouaynaya, “Robust explainability: A tutorial on gradient-based attribution methods for deep neural networks,”IEEE Signal Processing Magazine, vol. 39, no. 4, pp. 73–84, 2022.
[12] S. Salzberg, “A nearest hyperrectangle learning method,” Machine learning, vol. 6, pp. 251–276, 1991.
[13] J.-S. Jang, “Anfis: Adaptive-network-based fuzzy inference system,” IEEE transactions on systems, man, and cybernetics, vol. 23, no. 3, pp. 665–685, 1993.
[14] C.-T. Lin and C. S. G. Lee, “Neural-network-based fuzzy logic control and decision system,” IEEE Transactions on computers, vol. 40, no. 12, pp. 1320–1336, 1991.
[15] P. K. Simpson, “Fuzzy min-max neural networks. i. classification,” IEEE Trans. on Neural Networks, vol. 3, no. 5, pp. 776–786, 1992.
[16] M.-C. Su, “Use of neural networks as medical diagnosis expert systems,” Computers in biology and medicine, vol. 24, no. 6, pp. 419–429, 1994.
[17] Z. Yang, A. Zhang, and A. Sudjianto, “Enhancing explainability of neural networks through architecture constraints,” IEEE Transactions on Neural Networks and Learning Systems, vol. 32, no. 6, pp. 2610–2621, 2020.
[18] A. Sudjianto, Z. Yang, and A. Zhang, “Single-index model tree,” IEEE Transactions on Knowledge and Data Engineering, 2021.
[19] S. M. Lundberg and S.-I. Lee, “A unified approach to interpreting model predictions,”Advances in neural information processing systems, vol. 30, 2017.
[20] S. Bach, A. Binder, G. Montavon, F. Klauschen, K.-R. Müller, and W. Samek, “On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance propagation,” PloS one, vol. 10, no. 7, e0130140, 2015.
[21] K. Simonyan, A. Vedaldi, and A. Zisserman, “Deep inside convolutional networks: Visualising image classification models and saliency maps,” arXiv preprint arXiv:1312.6034, 2013.
[22] M. D. Zeiler and R. Fergus, “Visualizing and understanding convolutional networks,” in Computer Vision–ECCV 2014: 13th European Conference, Zurich, Switzerland, September 6-12, 2014, Proceedings, Part I 13, Springer, 2014, pp. 818–833.
[23] M. Sundararajan, A. Taly, and Q. Yan, “Axiomatic attribution for deep networks,” in International conference on machine learning, PMLR, 2017, pp. 3319–3328.
[24] D. Smilkov, N. Thorat, B. Kim, F. Viégas, and M. Wattenberg, “Smoothgrad: Removing noise by adding noise,” arXiv preprint arXiv:1706.03825, 2017.
[25] R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad-cam: Visual explanations from deep networks via gradient-based localization,” in Proceedings of the IEEE international conference on computer vision, 2017, pp. 618–626.
[26] M. Alber, S. Lapuschkin, P. Seegerer, et al., “Innvestigate neural networks!” J. Mach. Learn. Res., vol. 20, no. 93, pp. 1–8, 2019.
[27] M. T. Ribeiro, S. Singh, and C. Guestrin, “” why should i trust you?” explaining the predictions of any classifier,” in Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining, 2016, pp. 1135–1144.
[28] K. He, X. Zhang, S. Ren, and J. Sun, “Delving deep into rectifiers: Surpassing human-level performance on imagenet classification,” in Proceedings of the IEEE international conference on computer vision, 2015, pp. 1026–1034.
[29] H. Xiao, K. Rasul, and R. Vollgraf, “Fashion-mnist: A novel image dataset for benchmarking machine learning algorithms,” arXiv preprint arXiv:1708.07747, 2017.
[30] A. Krizhevsky, I. Sutskever, and G. E. Hinton, “Imagenet classification with deep convolutional neural networks,” Communications of the ACM, vol. 60, no. 6, pp. 84–90, 2017.
[31] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 770–778.
[32] C. Szegedy, W. Liu, Y. Jia, et al., “Going deeper with convolutions,” in Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 1–9.