Global Flood Prediction: a Multimodal Machine Learning Approach
This work addresses flood prediction for disaster management, but it is incremental as it applies existing multimodal techniques to a new domain.
The paper tackles global flood risk prediction by developing a multimodal machine learning approach that combines geographical and historical disaster data, achieving 75-77% ROCAUC scores for predicting flooding events up to 5 years ahead.
Flooding is one of the most destructive and costly natural disasters, and climate changes would further increase risks globally. This work presents a novel multimodal machine learning approach for multi-year global flood risk prediction, combining geographical information and historical natural disaster dataset. Our multimodal framework employs state-of-the-art processing techniques to extract embeddings from each data modality, including text-based geographical data and tabular-based time-series data. Experiments demonstrate that a multimodal approach, that is combining text and statistical data, outperforms a single-modality approach. Our most advanced architecture, employing embeddings extracted using transfer learning upon DistilBert model, achieves 75\%-77\% ROCAUC score in predicting the next 1-5 year flooding event in historically flooded locations. This work demonstrates the potentials of using machine learning for long-term planning in natural disaster management.