In the context of the increasing scale and complexity of disasters caused by rapid climate change, a comprehensive understanding of disaster big data is essential for effective detection and response. The disaster knowledge graph proposed in this paper fills this gap by capturing the connections between various disaster-related data sources and their potential for growth across heterogeneous datasets. We generate time-series disaster graphs every minute using SNS data (e.g., Twitter) and public data, specifically focusing on disasters. Then, we create disaster knowledge graphs to represent the relationships between various data sources and try to predict their potential developments. We label and annotate knowledge graphs and then detect sudden changes in time-series disaster knowledge graphs for disaster detection. To that end, we assess the effectiveness of three state-of-the-art GNN models for graph-based event classification using Graph Convolutional Network (GCN), Graph Attention Network (GAT), and SageConv. In addition, we evaluate a simple clustering model, K-means, for comparison. Our experiments show promising results with approximately 87% precision in detecting disaster events using structural data and connectivity patterns within disaster graphs. Finally, we measure the result of disaster detection time with an unseen dataset, showing positive results that about 70% detect a disaster in less than 3 minutes. To comprehensively analyze real-time social media data and understand the patterns of disaster to enhance disaster management and response strategies, our approach combines the strength of GNNs with a designed disaster knowledge graph.

Anomaly detection in temporal interaction networks (TINs) has become critical in network security, digital finance, and social networks. While recent studies based on Graph Neural Networks (GNNs) have yielded promising results, the existing methods are still limited by insufficient labels and noisy data, often ignoring the information filtering for unrelated user interactions. Therefore, this paper proposes a dynamic edge anomaly detection framework, AD-TIN, to address these challenges based on a multi-representation attention mechanism. It encodes graph structural information using a network information propagation module with neighbor sampling and graph diffusion. Furthermore, the network update module combines past node states with current structural features to capture the temporal information in potential user relationships, effectively mitigating the impact of noisy data. Extensive experiments on three real-world datasets demonstrate the robustness and efficacy of AD-TIN in addressing noise and unrelated interactions for edge anomaly detection.

The explosive rise of malware poses risks to Android developers and organization regarding security lapses and monetary losses. The dynamic nature, changing complexity and behavior over time, and increasing velocity and volume make it challenging for the malware protection community to provide a robust and reliable protection system. Due to these characteristics, conventional Android malware detection techniques, such as signature-based and battery-monitoring, cannot detect futuristic malware. Current research exploiting deep learning methods shows excellent performance compared to conventional and machine learning methods. However, the majority of the techniques are proposed for only binary classification. These classification models are tested on customized datasets. They do not provide the model’s effectiveness in terms of generalization, as the model’s accuracy might be good for some malware classes. Hence, providing a practical, robust, stable, and reliable malware model is still an open issue. Therefore, in this work, we propose an Attention-based deep learning model to detect categorical malware classes. The attention-based deep learning mechanism learns the malicious behavior of target classes. The attention mechanism filters and extracts the relevant information more effectively by focusing on the specific keywords in a target sample.

As a result of the recent surge in disaster-related data, numerous studies have been conducted to deal with the massive amount of data. In the meantime, the issue of managing data in various formats and representing their relevance is being raised. In this paper, we present a disaster knowledge graph to analyze the impact of a disaster and predict how much effort it will take to recover from the disaster. To that end, we define the structure of a disaster knowledge graph containing data collected from sensors, social networks, web, and risk analysis results. To extract meaningful information from structured and unstructured data, we use a risk analysis platform that can compute hazard values in accordance with various hazard models. Then, we store automatically graphs into a graph database as a form of a time-series data. Therefore, it will be possible to predict the progress of a complex disaster that can occur in a chain using a series of disaster knowledge graphs.

Due to the high population density and economic value of contemporary cities, earthquakes inflict greater damage on these cities. Consequently, the importance of quick earthquake early warning (EEW) is expanding, yet it is challenging to create a dense seismic monitoring network due to high installation and management costs. In order to overcome such limitations, MEMS sensors to monitor earthquakes and artificial intelligence (AI) technologies to analyze massive earthquake monitoring data are widely used today. In AI-based earthquake detection, the key to accurate detection is the use of sufficient data that accurately represents the various earthquake patterns. Unfortunately, how-ever, there is no publicly accessible database containing IoT-based seismic data. This is the result of relatively short research efforts. During the last two years of operation of CrowdQuake, a MEMS-based earthquake detection system, we collected earthquake and non-earthquake events, as well as normal noise data, which was greatly useful to improve the accuracy of AI models. As a result, we present an open dataset that is publicly available for MEMS-based earthquake detection research.

In recent years, a low-cost micro-electro-mechanical systems (MEMS) acceleration sensor has been widely used for earthquake early warning (EEW). In our previous work, we introduced a networked earthquake detection system, CrowdQuake with three-hundred smartphones’ acceleration sensors and a deep-learning based earthquake detection model. For one year’s operation, CrowdQuake detected a series of earthquakes and collected various earthquake and non-earthquake data. Based on the successful operation of CrowdQuake, in this paper, we discuss how it can be expanded across the country by addressing the following challenges: (1) sensor deployments for highly dense network, (2) earthquake detection performance using a deep learning model, and (3) high performance and scalable system design for big data processing. The improved system is CrowdQuake+ which can deal with acceleration data sent from 8,000 IoT sensors and detect an earthquake in few seconds using a newly proposed detection model. Moreover, CrowdQuake+ stores all acceleration data sent from sensors and assesses their qualities by calculating noise levels. Then, the collected data are used for deep learning model training, so that its detection performance becomes more accurate.

Recently, low-cost acceleration sensors have been widely used to detect earthquakes due to the significant development of MEMS technologies. It, however, still requires a high-density network to fully harness the low-cost sensors, especially for real-time earthquake detection. The design of a high-performance and scalable networked system thus becomes essential to be able to process a large amount of sensor data from hundreds to thousands of the sensors. An efficient and accurate earthquake-detection algorithm is also necessary to distinguish earthquake waveforms from various kinds of non-earthquake ones within the huge data in real time. In this paper, we present CrowdQuake, a networked system based on low-cost acceleration sensors, which monitors ground motions and detects earthquakes, by developing a convolutional-recurrent neural network model. This model ensures high detection performance while maintaining false alarms at a negligible level. We also provide detailed case studies on two of a few small earthquakes that have been detected by CrowdQuake during its last one-year operation.