Adib Hasan

VITA: Variational Pretraining of Transformers for Climate-Robust Crop Yield Forecasting

Accurate crop yield forecasting is essential for global food security. However, current AI models systematically underperform when yields deviate from historical trends. We attribute this to the lack of rich, physically grounded datasets directly linking atmospheric states to yields. To address this