Pairwise Coupling of Convolutional Neural Networks for the Better Explainability of Classification Systems
Source abstract
Abstract
Ensembling techniques are viewed as a promising machine learning tool to resolve issues arising from the monolithic nature of deep neural networks. In this paper, we consider pairwise coupling models built from neural networks, which is a special kind of ensemble. These models promise to provide much needed modularity in classification models employing deep networks. In order to be practical, pairwise coupling models have to have comparable memory and speed requirements to commonly used architectures. In this paper, we propose novel architectures that address this key problem of pairwise coupling models. We show that the classification accuracy of the resulting pairwise coupling models matches the original network, while exceeding the original network’s pairwise accuracy. The introduction of these pairwise models brings additional benefits. First, they allow for much less expensive uncertainty predictions. Secondly, their modularity allows for the fine-tuning of classification accuracy. Both of these benefits can be viewed as relating to the larger topic of improving the explainability as well as the modularity of deep neural networks.
Source: crossref · Quality: baseline
Research context
OpenAlex topics: Explainable Artificial Intelligence (XAI) · Adversarial Robustness in Machine Learning · Advanced Neural Network Applications
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
Affiliations
Mathematical Institute of Slovak Academy of Sciences, 974 11 Banská Bystrica, Slovakia
Faculty of management and informatics, University of Žilina, 010 26 Žilina, Slovakia
Funding
Ministry of Education, Science, Research and Sport of the Slovak Republic
References (42)
Show source-provided references
1. Deep Learning10.1038/nature14539 ↗
2. Explainable convolutional neural networks: A taxonomy, review, and future directions10.1145/3563691 ↗
3. A survey on neural network interpretability10.1109/TETCI.2021.3100641 ↗
4. An analysis of explainability methods for convolutional neural networks10.1016/j.engappai.2022.105606 ↗
5. Imagenet classification with deep convolutional neural networks
6. Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., and Torralba, A. (2016, January 27–30). Learning deep features for discriminative localization. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA.10.1109/CVPR.2016.319 ↗
7. Selvaraju, R.R., Das, A., Vedantam, R., Cogswell, M., Parikh, D., and Batra, D. (2016). Grad-CAM: Why did you say that?. arXiv.
8. Ribeiro, M.T., Singh, S., and Guestrin, C. (2016, January 13–17). “ Why should I trust you?” Explaining the predictions of any classifier. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA.10.1145/2939672.2939778 ↗
9. Zeiler, M.D., and Fergus, R. (2014, January 6–12). Visualizing and understanding convolutional networks. Proceedings of the European Conference on Computer Vision, Zurich, Switzerland.10.1007/978-3-319-10590-1_53 ↗
10. Dietterich, T.G. (2000, January 21–23). Ensemble methods in machine learning. Proceedings of the International Workshop on Multiple Classifier Systems, Cagliari, Italy.10.1007/3-540-45014-9_1 ↗
11. Modularity in neural computing10.1109/5.784227 ↗
12. Adaptive Mixtures of Local Experts10.1162/neco.1991.3.1.79 ↗
13. Liu, A., Feng, B., Xue, B., Wang, B., Wu, B., Lu, C., Zhao, C., Deng, C., Zhang, C., and Ruan, C. (2024). Deepseek-v3 technical report. arXiv.
14. Explainability of deep neural networks for MRI analysis of brain tumors10.1007/s11548-022-02619-x ↗
15. A novel explainable neural network for Alzheimer’s disease diagnosis10.1016/j.patcog.2022.108876 ↗
16. Neural network-based uncertainty quantification: A survey of methodologies and applications10.1109/ACCESS.2018.2836917 ↗
17. A survey of uncertainty in deep neural networks10.1007/s10462-023-10562-9 ↗
18. James, G., Witten, D., Hastie, T., and Tibshirani, R. (2014). An Introduction to Statistical Learning: With Applications in R, Springer Publishing Company.10.1007/978-1-4614-7138-7 ↗
19. Support-vector networks10.1023/A:1022627411411 ↗
20. Vapnik, V. (1996). The Nature of Statistical Learning Theory, Springer.10.1007/978-1-4757-2440-0 ↗
21. Dorffner, G., Bischof, H., and Hornik, K. (2001, January 21–25). Multiclass Classification with Pairwise Coupled Neural Networks or Support Vector Machines. Proceedings of the Artificial Neural Networks—ICANN 2001, Vienna, Austria.10.1007/3-540-44668-0 ↗
22. Platt, J.C. (1998, January 1–3). Probabilies for SV machines. Proceedings of the Advances in Large Margin Classifiers, Breckenridge, CO, USA.
23. Classification by pairwise coupling10.1214/aos/1028144844 ↗
24. Classification by pairwise coupling of imprecise probabilities10.1016/j.patcog.2017.10.019 ↗
25. Šuch, O., Beňuš, Š., and Tinajová, A. (2015, January 17–21). A new method to combine probability estimates from pairwise binary classifiers. Proceedings of the Information Technologies–Applications and Theory, ITAT 2015, Slovensky Raj, Slovakia.
26. Tesauro, G., Touretzky, D., and Leen, T. (1995). Pairwise Neural Network Classifiers with Probabilistic Outputs. Advances in Neural Information Processing Systems 7, MIT Press.
27. Probability Estimates for Multi-class Classification by Pairwise Coupling
28. Bayes covariant multi-class classification10.1016/j.patrec.2016.08.014 ↗
29. LIBSVM: A Library for Support Vector Machines10.1145/1961189.1961199 ↗
30. A partitioned neural network approach for vowel classification using smoothed time/frequency features10.1109/89.771263 ↗
31. Xiao, H., Rasul, K., and Vollgraf, R. (2017). Fashion-MNIST: A Novel Image Dataset for Benchmarking Machine Learning Algorithms. arXiv.
32. LeCun, Y., and Cortes, C. (2019, April 01). MNIST Handwritten Digit Database. Available online: http://yann.lecun.com/exdb/mnist/.
33. Strubell, E., Ganesh, A., and McCallum, A. (2019). Energy and Policy Considerations for Deep Learning in NLP. arXiv.10.18653/v1/P19-1355 ↗
34. Labach, A., Salehinejad, H., and Valaee, S. (2019). Survey of Dropout Methods for Deep Neural Networks. arXiv.
35. Bach, F., and Blei, D. (2015, January 7–9). Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. Proceedings of the 32nd International Conference on Machine Learning, Lille, France.
36. He, K., Zhang, X., Ren, S., and Sun, J. (2016, January 27–30). Deep Residual Learning for Image Recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA.10.1109/CVPR.2016.90 ↗
37. Šuch, O., Tinajová, A., and Kontšek, M. (2019, January 20–24). Neural pairwise classification models created by ignoring irrelevant alternatives. Proceedings of the ITAT 2015: Information Technologies–Applications and Theory, Donovaly, Slovakia.
38. Complementary Log Regression for Generalized Linear Models10.1080/00031305.1992.10475858 ↗
39. Lakshminarayanan, B., Pritzel, A., and Blundell, C. (2017, January 4–9). Simple and scalable predictive uncertainty estimation using deep ensembles. Proceedings of the Advances in Neural Information Processing Systems NIPS’17, Long Beach, CA, USA.
40. Gal, Y., and Ghahramani, Z. (2016, January 19–24). Dropout as a bayesian approximation: Representing model uncertainty in deep learning. Proceedings of the International Conference on Machine Learning, PMLR, New York, NY, USA.
41. Kendall, A., and Gal, Y. (2017, January 4–9). What uncertainties do we need in bayesian deep learning for computer vision?. Proceedings of the Advances in Neural Information Processing Systems NIPS’17, Long Beach, CA, USA.
42. Šuch, O., Novotný, P., and Haidar, A. (J. Classif., 2025). New Bayes covariant coupling methods for adapting to change of priors in one-vs-one classification, J. Classif., under review.