File Exchange

image thumbnail

Explainable AI: interpreting the classification using LIME

version 2.0.3 (2 MB) by Kenta
This demo shows how to interpret the classification by CNN using LIME (Local Interpretable Model-agnostic Explanations) [1]. LIMEによる特徴量の可視化

20 Downloads

Updated 26 Oct 2020

View Version History

View License

[English]
This demo shows how to interpret the classification by CNN using LIME (Local Interpretable Model-agnostic Explanations) [1]. This demo was created based on [1], but the implementation might be a little bit different from its official one. This code highlights the regions that contributed to the classification. It helps you interpret and improve the model, or, you can recoginize the classifier is not untrustworthy for you if the region highlighted is irrelevant for the true class.
This demo shows the example with a pre-trained model of ResNet-18 [2]. Please modify this code based on your network. LIME can be utilized for any type of data and this example shows the code for image classification.

[Japanese]
LIME [1]と呼ばれる手法を用いて、深層学習による画像分類を行った際の重要箇所を可視化します。公式の実装とは若干の違いがあるかもしれませんのでご注意ください。サムネイルでは、学習済みネットワーク(ResNet-18 [2])が、ゴールデンレトリバーと予測したときの判断根拠の位置を可視化しています。

[Key words]
classification, cnn (convolutional neural network), deep learning, explainable AI, image, interpret, LIME (Local Interpretable Model-agnostic Explanations), machine learning, superpixel, visualization, why

[1] Ribeiro, M.T., Singh, S. and Guestrin, C., 2016, August. " Why should I trust you?" Explaining the predictions of any classifier. In Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining (pp. 1135-1144).
[2] He, K., Zhang, X., Ren, S. and Sun, J., 2016. Deep residual learning for image recognition. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 770-778).

Cite As

Kenta (2020). Explainable AI: interpreting the classification using LIME (https://www.mathworks.com/matlabcentral/fileexchange/77828-explainable-ai-interpreting-the-classification-using-lime), MATLAB Central File Exchange. Retrieved .

Comments and Ratings (5)

雄二朗 安井

Takuji Fukumoto

Tohru Kikawada

michio

Yoshito Saito

MATLAB Release Compatibility
Created with R2020a
Compatible with any release
Platform Compatibility
Windows macOS Linux

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!