Disaster Detection on Social Media Using Deep Multimodal Attention Models

The deep attentive multimodal learning technique to catastrophe identification from social media posts is introduced in this paper. Convolutional neural networks (CNNs) are used in the proposed model to integrate text and image data, while bidirectional long short-term memory (BiLSTM) networks with attention mechanisms are used for text analysis. In an effort to improve catastrophe detection accuracy and give emergency responders and decision-makers timely information, the model makes use of multimodal data fusion approaches. According to experimental data, the strategy performs better than conventional techniques and provides a reliable means of identifying and monitoring disasters in real time on social media platforms.