AI & ML

Vegetation Risk Analysis (GAT)

Grid Attention Transformer (GAT) is a transformer-based model for spatio-temporal vegetation risk analysis using multi-year remote sensing data.

Back to Research
Vegetation Risk Analysis (GAT) project illustration.

Research Area

AI & ML

Key Methods

Transformers, Remote Sensing, Spatio-Temporal Modeling

Project Summary

Grid attention transformer for spatio-temporal vegetation forecasting and risk analysis.

Environmental risk analysis requires models that can capture meaningful vegetation change patterns across both space and time rather than only snapshot-based signals. Model long-range spatio-temporal vegetation change and support environmental forecasting and risk assessment. The project tracks change patterns across regions such as Anyinam and Serikoberi over the 2017 to 2023 period. It uses a transformer architecture tailored to grid-based spatio-temporal analysis and remote sensing workflows. The main objective is to support vegetation risk analysis with stronger temporal awareness and better regional context.