Skip to main content

2021 | OriginalPaper | Buchkapitel

Causal Explanation of Convolutional Neural Networks

Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.

search-config
loading …

Abstract

In this paper we introduce an explanation technique for Convolutional Neural Networks (CNNs) based on the theory of causality by Halpern and Pearl [12]. The causal explanation technique (CexCNN) is based on measuring the filter importance to a CNN decision, which is measured through counterfactual reasoning. In addition, we employ extended definitions of causality, which are responsibility and blame to weight the importance of such filters and project their contribution on input images. Since CNNs form a hierarchical structure, and since causal models can be hierarchically abstracted, we employ this similarity to perform the most important contribution of this paper, which is localizing the important features in the input image that contributed the most to a CNN’s decision. In addition to its ability in localization, we will show that CexCNN can be useful as well for model compression through pruning the less important filters. We tested CexCNN on several CNNs architectures and datasets. (The code is available on https://​github.​com/​HichemDebbi/​CexCNN)

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

  • über 102.000 Bücher
  • über 537 Zeitschriften

aus folgenden Fachgebieten:

  • Automobil + Motoren
  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Elektrotechnik + Elektronik
  • Energie + Nachhaltigkeit
  • Finance + Banking
  • Management + Führung
  • Marketing + Vertrieb
  • Maschinenbau + Werkstoffe
  • Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Springer Professional "Technik"

Online-Abonnement

Mit Springer Professional "Technik" erhalten Sie Zugriff auf:

  • über 67.000 Bücher
  • über 390 Zeitschriften

aus folgenden Fachgebieten:

  • Automobil + Motoren
  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Elektrotechnik + Elektronik
  • Energie + Nachhaltigkeit
  • Maschinenbau + Werkstoffe




 

Jetzt Wissensvorsprung sichern!

Springer Professional "Wirtschaft"

Online-Abonnement

Mit Springer Professional "Wirtschaft" erhalten Sie Zugriff auf:

  • über 67.000 Bücher
  • über 340 Zeitschriften

aus folgenden Fachgebieten:

  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Finance + Banking
  • Management + Führung
  • Marketing + Vertrieb
  • Versicherung + Risiko




Jetzt Wissensvorsprung sichern!

Literatur
4.
Zurück zum Zitat Avanti, S., Peyton, G., Anshul, K.: Learning important features through propagating activation differences, pp. 3145–3153. ICML’17 (2017) Avanti, S., Peyton, G., Anshul, K.: Learning important features through propagating activation differences, pp. 3145–3153. ICML’17 (2017)
5.
Zurück zum Zitat Beckers, S., Halpern, J.Y.: Abstracting causal models. In: AAAI (2017) Beckers, S., Halpern, J.Y.: Abstracting causal models. In: AAAI (2017)
6.
Zurück zum Zitat Chockler, H., Halpern, J.Y.: Responsibility and blame: a structural-model approach. J. Artif. Int. Res. 22(1), 93–115 (2004)MathSciNetMATH Chockler, H., Halpern, J.Y.: Responsibility and blame: a structural-model approach. J. Artif. Int. Res. 22(1), 93–115 (2004)MathSciNetMATH
7.
Zurück zum Zitat Choe, J., Oh, S.J., Lee, S., Chun, S., Akata, Z., Shim, H.: Evaluating weakly supervised object localization methods right. In: CVPR, pp. 3130–3139 (2020) Choe, J., Oh, S.J., Lee, S., Chun, S., Akata, Z., Shim, H.: Evaluating weakly supervised object localization methods right. In: CVPR, pp. 3130–3139 (2020)
8.
Zurück zum Zitat Gordon, D., Kembhavi, A., Rastegari, M., Redmon, J., Fox, D., Farhadi, A.: Iqa: visual question answering in interactive environments. In: In arXiv:1712.03316 (2017) Gordon, D., Kembhavi, A., Rastegari, M., Redmon, J., Fox, D., Farhadi, A.: Iqa: visual question answering in interactive environments. In: In arXiv:​1712.​03316 (2017)
9.
Zurück zum Zitat Deng, J., Dong, W., Socher, R., Li, L., Li, K., Fei-Fei, L.: Imagenet: a large-scale hierarchical image database. In: CVPR, pp. 248–255 (2009) Deng, J., Dong, W., Socher, R., Li, L., Li, K., Fei-Fei, L.: Imagenet: a large-scale hierarchical image database. In: CVPR, pp. 248–255 (2009)
10.
Zurück zum Zitat Girshick, R., Donahue, J., Darrell, T., Malik, J.: Rich feature hierarchies for accurate object detection and semantic segmentation. In: CVPR, pp. 580–587 (2014) Girshick, R., Donahue, J., Darrell, T., Malik, J.: Rich feature hierarchies for accurate object detection and semantic segmentation. In: CVPR, pp. 580–587 (2014)
11.
Zurück zum Zitat Goodfellow, I.J., Shlens, J., Szegedy, C.: Explaining and harnessing adversarial examples. In: ICLR (2015) Goodfellow, I.J., Shlens, J., Szegedy, C.: Explaining and harnessing adversarial examples. In: ICLR (2015)
12.
Zurück zum Zitat Halpern, J., Pearl, J.: Causes and explanations: a structural-model approach part i: Causes. In: Proceedings of the 17th UAI, pp. 194–202 (2001) Halpern, J., Pearl, J.: Causes and explanations: a structural-model approach part i: Causes. In: Proceedings of the 17th UAI, pp. 194–202 (2001)
13.
Zurück zum Zitat Halpern, J.Y., Pearl, J.: Causes and explanations: a structural-model approach. part ii: Explanations. Br. J. Philos. Sci. 56(4), 889–911 (2008) Halpern, J.Y., Pearl, J.: Causes and explanations: a structural-model approach. part ii: Explanations. Br. J. Philos. Sci. 56(4), 889–911 (2008)
14.
Zurück zum Zitat Han, S., Pool, J., Tran, J., Dally, W.J.: Learning both weights and connections for efficient neural networks. In: NIPS (2015) Han, S., Pool, J., Tran, J., Dally, W.J.: Learning both weights and connections for efficient neural networks. In: NIPS (2015)
15.
Zurück zum Zitat Harradon, M., Druce, J., Ruttenberg, B.E.: Causal learning and explanation of deep neural networks via autoencoded activations. In: CoRR abs/1802.00541 (2018) Harradon, M., Druce, J., Ruttenberg, B.E.: Causal learning and explanation of deep neural networks via autoencoded activations. In: CoRR abs/1802.00541 (2018)
16.
Zurück zum Zitat Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep convolutional neural networks. Commun. ACM 60(6), 84–90 (2017)CrossRef Krizhevsky, A., Sutskever, I., Hinton, G.E.: Imagenet classification with deep convolutional neural networks. Commun. ACM 60(6), 84–90 (2017)CrossRef
17.
Zurück zum Zitat Lecun, Y., Bottou, L., Bengio, Y., Haffner, P.: Gradient-based learning applied to document recognition. Proc. IEEE 86(11), 2278–2324 (1998)CrossRef Lecun, Y., Bottou, L., Bengio, Y., Haffner, P.: Gradient-based learning applied to document recognition. Proc. IEEE 86(11), 2278–2324 (1998)CrossRef
20.
Zurück zum Zitat Li, H., Kadav, A., Durdanovic, I., Samety, H.: Pruning filters for efficient convnets. In: ICLR 2017, pp. 1–13 (2017) Li, H., Kadav, A., Durdanovic, I., Samety, H.: Pruning filters for efficient convnets. In: ICLR 2017, pp. 1–13 (2017)
21.
Zurück zum Zitat Long, J., Shelhamer, E., Darrell, T.: Fully convolutional networks for semantic segmentation. In: CVPR, pp. 3431–3440 (2015) Long, J., Shelhamer, E., Darrell, T.: Fully convolutional networks for semantic segmentation. In: CVPR, pp. 3431–3440 (2015)
22.
Zurück zum Zitat Lundberg, S.M., Lee, S.I.: A unified approach to interpreting model predictions. In: NIPS, pp. 4768–4777 (2017) Lundberg, S.M., Lee, S.I.: A unified approach to interpreting model predictions. In: NIPS, pp. 4768–4777 (2017)
24.
Zurück zum Zitat Narendra, T., Sankaran, A., Vijaykeerthy, D., Mani, S.: Explaining deep learning models using causal inference. In: arXiv:1811.04376 (2018) Narendra, T., Sankaran, A., Vijaykeerthy, D., Mani, S.: Explaining deep learning models using causal inference. In: arXiv:​1811.​04376 (2018)
25.
Zurück zum Zitat Ribeiro, M.T., Singh, S., Guestrin, C.: “why should i trust you?": Explaining the predictions of any classifier, pp. 1135–1144. KDD ’16 (2016) Ribeiro, M.T., Singh, S., Guestrin, C.: “why should i trust you?": Explaining the predictions of any classifier, pp. 1135–1144. KDD ’16 (2016)
26.
Zurück zum Zitat Russakovsky, O., et al.: Imagenet large scale visual recognition challenge. Int. J. Comput. Vis. 115(3), 211–252 (2015)MathSciNetCrossRef Russakovsky, O., et al.: Imagenet large scale visual recognition challenge. Int. J. Comput. Vis. 115(3), 211–252 (2015)MathSciNetCrossRef
27.
Zurück zum Zitat Schwab, P., Karlen, W.: Cxplain: causal explanations for model interpretation under uncertainty. NeurIPS, pp. 10220–10230 (2019) Schwab, P., Karlen, W.: Cxplain: causal explanations for model interpretation under uncertainty. NeurIPS, pp. 10220–10230 (2019)
28.
Zurück zum Zitat Selvaraju, R.R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., Batra, D.: Grad-cam: visual explanations from deep networks via gradient-based localization. In: ICCV, pp. 618–626 (2017) Selvaraju, R.R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., Batra, D.: Grad-cam: visual explanations from deep networks via gradient-based localization. In: ICCV, pp. 618–626 (2017)
29.
Zurück zum Zitat Simonyan, K., Vedaldi, A., Zisserman, A.: Deep inside convolutional networks: visualising image classification models and saliency maps. In: In arXiv:1312.6034 (2013) Simonyan, K., Vedaldi, A., Zisserman, A.: Deep inside convolutional networks: visualising image classification models and saliency maps. In: In arXiv:​1312.​6034 (2013)
30.
Zurück zum Zitat Singh, K.K., Lee, Y.J.: Forcing a network to be meticulous for weakly-supervised object and action localization. In: CVPR (2017) Singh, K.K., Lee, Y.J.: Forcing a network to be meticulous for weakly-supervised object and action localization. In: CVPR (2017)
31.
Zurück zum Zitat Smilkov, D., Thorat, N., Kim, B., Viegas, F.B., Wattenberg, M.: Smoothgrad: removing noise by adding noise. In: CoRR, vol. abs/1706.03825 (2017) Smilkov, D., Thorat, N., Kim, B., Viegas, F.B., Wattenberg, M.: Smoothgrad: removing noise by adding noise. In: CoRR, vol. abs/1706.03825 (2017)
32.
Zurück zum Zitat Sundararajan, M., Taly, A., Yan, Q.: Axiomatic attribution for deep networks. In: ICML, pp. 3319–3328 (2017) Sundararajan, M., Taly, A., Yan, Q.: Axiomatic attribution for deep networks. In: ICML, pp. 3319–3328 (2017)
34.
Zurück zum Zitat Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., Torralba, A.: Learning deep features for discriminative localization. In: CVPR, pp. 2921–2929 (2016) Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., Torralba, A.: Learning deep features for discriminative localization. In: CVPR, pp. 2921–2929 (2016)
Metadaten
Titel
Causal Explanation of Convolutional Neural Networks
verfasst von
Hichem Debbi
Copyright-Jahr
2021
DOI
https://doi.org/10.1007/978-3-030-86520-7_39

Premium Partner