TY - GEN
T1 - Explaining convolutional neural networks using softmax gradient layer-wise relevance propagation
AU - Iwana, Brian Kenji
AU - Kuroki, Ryohei
AU - Uchida, Seiichi
PY - 2019/10
Y1 - 2019/10
N2 - Convolutional Neural Networks (CNN) have become state-of-the-art in the field of image classification. However, not everything is understood about their inner representations. This paper tackles the interpretability and explainability of the predictions of CNNs for multi-class classification problems. Specifically, we propose a novel visualization method of pixel-wise input attribution called Softmax-Gradient Layer-wise Relevance Propagation (SGLRP). The proposed model is a class discriminate extension to Deep Taylor Decomposition (DTD) using the gradient of softmax to back propagate the relevance of the output probability to the input image. Through qualitative and quantitative analysis, we demonstrate that SGLRP can successfully localize and attribute the regions on input images which contribute to a target object's classification. We show that the proposed method excels at discriminating the target objects class from the other possible objects in the images. We confirm that SGLRP performs better than existing Layer-wise Relevance Propagation (LRP) based methods and can help in the understanding of the decision process of CNNs.
AB - Convolutional Neural Networks (CNN) have become state-of-the-art in the field of image classification. However, not everything is understood about their inner representations. This paper tackles the interpretability and explainability of the predictions of CNNs for multi-class classification problems. Specifically, we propose a novel visualization method of pixel-wise input attribution called Softmax-Gradient Layer-wise Relevance Propagation (SGLRP). The proposed model is a class discriminate extension to Deep Taylor Decomposition (DTD) using the gradient of softmax to back propagate the relevance of the output probability to the input image. Through qualitative and quantitative analysis, we demonstrate that SGLRP can successfully localize and attribute the regions on input images which contribute to a target object's classification. We show that the proposed method excels at discriminating the target objects class from the other possible objects in the images. We confirm that SGLRP performs better than existing Layer-wise Relevance Propagation (LRP) based methods and can help in the understanding of the decision process of CNNs.
UR - http://www.scopus.com/inward/record.url?scp=85082458309&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85082458309&partnerID=8YFLogxK
U2 - 10.1109/ICCVW.2019.00513
DO - 10.1109/ICCVW.2019.00513
M3 - Conference contribution
T3 - Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019
SP - 4176
EP - 4185
BT - Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 17th IEEE/CVF International Conference on Computer Vision Workshop, ICCVW 2019
Y2 - 27 October 2019 through 28 October 2019
ER -