GraphVSSM: Graph Variational State-Space Model for Probabilistic Spatiotemporal Inference of Dynamic Exposure and Vulnerability for Regional Disaster Resilience Assessment

2025-08-02

AI Summary

Current regional disaster resilience assessments face persistent challenges in physical vulnerability modeling—including static assumptions, coarse spatial granularity, insufficient calibration, and poor generalizability. To address these limitations, we propose GraphVSSM: the first weakly supervised probabilistic spatiotemporal inference framework integrating graph neural networks, state-space models, and variational inference. Leveraging time-series satellite imagery and expert priors, GraphVSSM enables interpretable, city-scale vulnerability modeling—from coarse to fine resolution—while explicitly capturing spatial dependencies, temporal dynamics, and epistemic uncertainty. It overcomes the constraints of conventional static aggregation. We validate the method across the Philippines, Bangladesh, and Sierra Leone, and release METEOR 2.5D, an enhanced open dataset. This advancement significantly improves high-resolution disaster risk monitoring and evidence-based resilience policy support in least-developed countries.

Technology Category

• Reasoning under Uncertainty: Graphical Models
• Computer Vision: Remote Sensing / Geospatial AI
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

Regional disaster resilience quantifies the changing nature of physical risks to inform policy instruments ranging from local immediate recovery to international sustainable development. While many existing state-of-practice methods have greatly advanced the dynamic mapping of exposure and hazard, our understanding of large-scale physical vulnerability has remained static, costly, limited, region-specific, coarse-grained, overly aggregated, and inadequately calibrated. With the significant growth in the availability of time-series satellite imagery and derived products for exposure and hazard, we focus our work on the equally important yet challenging element of the risk equation: physical vulnerability. We leverage machine learning methods that flexibly capture spatial contextual relationships, limited temporal observations, and uncertainty in a unified probabilistic spatiotemporal inference framework. We therefore introduce Graph Variational State-Space Model (GraphVSSM), a novel modular spatiotemporal approach that uniquely integrates graph deep learning, state-space modeling, and variational inference using time-series data and prior expert belief systems in a weakly supervised or coarse-to-fine-grained manner. We present three major results: a city-wide demonstration in Quezon City, Philippines; an investigation of sudden changes in the cyclone-impacted coastal Khurushkul community (Bangladesh) and mudslide-affected Freetown (Sierra Leone); and an open geospatial dataset, METEOR 2.5D, that spatiotemporally enhances the existing global static dataset for UN Least Developed Countries (2020). Beyond advancing regional disaster resilience assessment and improving our understanding global disaster risk reduction progress, our method also offers a probabilistic deep learning approach, contributing to broader urban studies that require compositional data analysis in weak supervision.

Problem

• Dynamic mapping of physical vulnerability for disaster resilience
• Probabilistic spatiotemporal inference using limited temporal observations
• Integration of graph deep learning and variational inference

Innovation

• Graph deep learning for spatial contextual relationships
• State-space modeling for temporal observations
• Variational inference for uncertainty handling