Explanation of Multi-Label Neural Networks with Layer-Wise Relevance Propagation
Bello,Marilyn Napoles,Gonzalo Vanhoof,Koen Garcia,Maria M. Bello,Rafael
Bello,Marilyn
Napoles,Gonzalo
Vanhoof,Koen
Garcia,Maria M.
Bello,Rafael
Abstract
Neural networks are considered a black-box model as their strength in modeling complex interactions makes its operation almost impossible to explain. Still, neural networks remain very interesting tools as they have shown promising performance in various classification tasks. Layer-wise relevance propagation is a technique that, based on a propagation approach, is able to explain the predictions obtained by a neural network. In this work, we propose four adaptations of this technique to operate on multi-label neural networks. The proposed methods provide new ways of distributing the relevance between the output layer and the preceding ones. The efficacy of these adaptations is demonstrated after an experimental study. The study is carried out based on existing evaluation criteria in the literature that measure the explanation's quality. These methods are applied to a case study in which a neural network is used to detect secondary coinfections in patients infected with SARS-CoV-2. Overall, the proposed methods provide a post-hoc interpretability stage of the results.
Description
Funding Information: The authors would like to sincerely thank the doctors of the Hospital “Comandante Manuel Fajardo Rivero” in the city of Santa Clara, Cuba, who assisted us in the description and medical terminology associated with the case study under consideration. Likewise, in the validation of the results obtained as an expert in the field. This study is supported by the Special Research Fund of Hasselt University. Publisher Copyright: © 2022 IEEE.
Date
2022
Journal Title
Journal ISSN
Volume Title
Publisher
Institute of Electrical and Electronics Engineers Inc.
Research Projects
Organizational Units
Journal Issue
Keywords
Explanation, Layer-wise Relevance Propagation, Multi-label Scenarios, Neural Networks, SDG 3 - Good Health and Well-being
Citation
Bello, M, Napoles, G, Vanhoof, K, Garcia, M M & Bello, R 2022, Explanation of Multi-Label Neural Networks with Layer-Wise Relevance Propagation. in 2022 International Joint Conference on Neural Networks, IJCNN 2022 - Proceedings. Proceedings of the International Joint Conference on Neural Networks, vol. 2022-July, Institute of Electrical and Electronics Engineers Inc., 2022 International Joint Conference on Neural Networks, IJCNN 2022, Padua, Italy, 18/07/22. https://doi.org/10.1109/IJCNN55064.2022.9892239
License
info:eu-repo/semantics/openAccess