Earthquake-Induced Building-Damage Mapping Using Explainable AI (XAI)

Abstract

Building-damage mapping using remote sensing images plays a critical role in providing quick and accurate information for the first responders after major earthquakes. In recent years, there has been an increasing interest in generating post-earthquake building-damage maps automatically using different artificial intelligence (AI)-based frameworks. These frameworks in this domain are promising, yet not reliable for several reasons, including but not limited to the site-specific design of the methods, the lack of transparency in the AI-model, the lack of quality in the labelled image, and the use of irrelevant descriptor features in building the AI-model. Using explainable AI (XAI) can lead us to gain insight into identifying these limitations and therefore, to modify the training dataset and the model accordingly. This paper proposes the use of SHAP (Shapley additive explanation) to interpret the outputs of a multilayer perceptron (MLP)—a machine learning model—and analyse the impact of each feature descriptor included in the model for building-damage assessment to examine the reliability of the model. In this study, a post-event satellite image from the 2018 Palu earthquake was used. The results show that MLP can classify the collapsed and non-collapsed buildings with an overall accuracy of 84% after removing the redundant features. Further, spectral features are found to be more important than texture features in distinguishing the collapsed and non-collapsed buildings. Finally, we argue that constructing an explainable model would help to understand the model’s decision to classify the buildings as collapsed and non-collapsed and open avenues to build a transferable AI model.

The challenges of integrating explainable artificial intelligence into GeoAI

“…Frisoni et al (2020) argued that plain text explanations have been insufficiently formalized so they used mapping to visualize the position of given keywords in semantic space. Matin and Pradhan (2021) employed SHAP values to explain, via a beeswarm plot, earthquake features within DNN. Even as methods are used in the explanation and interpretation of GeoAI (Ma et al, 2021), most visualization explanations are not georeferenced.…”

“…Well-cited XAI visualization tools (e.g., https://github.com/yosin ski/deep-visua lizat ion-toolbox) display images as ideal types but lose geographic context like georeferences. In Section 2, we briefly described Matin and Pradhan's (2021) research on XAI to explain earthquake damage. XAI and DNN both extract common features (here, image patches) of all earthquake damaged buildings at any given layer.…”

Improve the Deep Learning Models in Forestry Based on Explanations and Expertise

“…Several studies use multiple XAI methods to interpret trained models, such as using Grad-CAM to obtain the contributions of input pixels as well as visualizing the feature maps of part layers (Xing et al, 2020 ). In addition to computer science, XAI methods have been applied in various fields including medicine (Tjoa and Guan, 2020 ), geography (Cheng et al, 2021 ), and disaster assessment (Matin and Pradhan, 2021 ). However, few studies have attempted to interpret models in the field of forestry (Onishi and Ise, 2021 ), even though deep learning methods have been widely applied in this field.…”

Global Spatial Suitability Mapping of Wind and Solar Systems Using an Explainable AI-Based Approach

“…XAI refers to AI algorithms in which humans can logically interpret the output at an acceptable level [24]. Explainability facilitates understanding the influence and contribution of each input feature on the AI models' outputs [25]. Furthermore, XAI has the ability to detect bias in the training dataset, thereby ensuring impartiality in decision-making [26].…”

“…Furthermore, XAI has the ability to detect bias in the training dataset, thereby ensuring impartiality in decision-making [26]. XAI has been increasingly applied in the field of spatial prediction of droughts [27,28] and landslides [29], mapping of earthquake-induced building damage [25], and urban vegetation mapping [30]. Most of these studies concluded that XAI can not only provide insight into the output of intelligent models but also change our understanding of using MLbased models to make informed decisions [28].…”