• shareshare
  • link
  • cite
  • add
auto_awesome_motion View all 4 versions
Publication . Conference object . Part of book or chapter of book . 2018

A Rule Extraction Study Based on a Convolutional Neural Network

Guido Bologna;
Open Access
Published: 27 Aug 2018
Publisher: HAL CCSD

Part 5: MAKE Explainable AI; International audience; Convolutional Neural Networks (CNNs) lack an explanation capability in the form of propositional rules. In this work we define a simple CNN architecture having a unique convolutional layer, then a Max-Pool layer followed by a full connected layer. Rule extraction is performed after the Max-Pool layer with the use of the Discretized Interpretable Multi Layer Perceptron (DIMLP). The antecedents of the extracted rules represent responses of convolutional filters, which are difficult to understand. However, we show in a sentiment analysis problem that from these “meaningless” values it is possible to obtain rules that represent relevant words in the antecedents. The experiments illustrate several examples of rules that represent n-grams.

Subjects by Vocabulary

Microsoft Academic Graph classification: Layer (object-oriented design) Discretization Multilayer perceptron Convolutional neural network Sentiment analysis Artificial intelligence business.industry business Pattern recognition Simple (abstract algebra) Computer science


Convolutional Neural Networks, Rule extraction, Sentiment analysis, [INFO]Computer Science [cs], [SHS.INFO]Humanities and Social Sciences/Library and information sciences, [INFO] Computer Science [cs], [SHS.INFO] Humanities and Social Sciences/Library and information sciences

14 references, page 1 of 2

1. Andrews, R., Diederich, J., Tickle, A.B.: Survey and critique of techniques for extracting rules from trained arti cial neural networks. Knowledge-based systems 8(6), 373{389 (1995) [OpenAIRE]

2. Bologna, G.: Rule extraction from a multilayer perceptron with staircase activation functions. In: Neural Networks, 2000. IJCNN 2000, Proceedings of the IEEE-INNSENNS International Joint Conference on. vol. 3, pp. 419{424. IEEE (2000)

3. Bologna, G.: A model for single and multiple knowledge based networks. Arti cial Intelligence in Medicine 28(2), 141{163 (2003)

4. Bologna, G., Hayashi, Y.: A rule extraction study from svm on sentiment analysis. Big Data and Cognitive Computing 2(1), 6 (2018)

5. Cliche, M.: Bb twtr at semeval-2017 task 4: Twitter sentiment analysis with cnns and lstms. arXiv preprint arXiv:1704.06125 (2017) [OpenAIRE]

6. Diederich, J., Dillon, D.: Sentiment recognition by rule extraction from support vector machines. In: CGAT 09 Proceedings: Computer Games, Multimedia and Allied Technology 09. Global Science and Technology Forum (2009)

7. Dieleman, S., Schlter, J., Raffel, C., Olson, E., Snderby, S.K., Nouri, D., et al.: Lasagne: First release. (Aug 2015).,

8. Holzinger, A., Biemann, C., Pattichis, C.S., Kell, D.B.: What do we need to build explainable ai systems for the medical domain? arXiv preprint arXiv:1712.09923 (2017) [OpenAIRE]

9. Kim, Y.: Convolutional neural networks for sentence classi cation. arXiv preprint arXiv:1408.5882 (2014)

10. Koh, P.W., Liang, P.: Understanding black-box predictions via in uence functions. arXiv preprint arXiv:1703.04730 (2017)

Related to Research communities
Download fromView all 3 sources
Part of book or chapter of book
License: cc-by
Providers: UnpayWall