File size: 1,360 Bytes
5617f12 d948c64 bfb8010 d948c64 bfb8010 b82f052 bfb8010 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 | ---
license: mit
datasets:
- notadib/NASA-Power-Daily-Weather
language:
- en
metrics:
- r_squared
tags:
- weather,
- agriculture,
---
# VITA: Variational Pretraining of Transformers for Climate-Robust Crop Yield Forecasting
This is the official pretrained model weights for the paper [arXiv:2508.03589](https://arxiv.org/abs/2508.03589). VITA is a variational pretraining framework that learns weather representations from rich satellite data and transfers them to yield prediction tasks with limited ground-based measurements.
## Overview
VITA addresses the data asymmetry problem in agricultural AI: pretraining uses 31 meteorological variables from NASA POWER satellite data, while deployment relies on only 6 basic weather features. Through variational pretraining with a seasonality-aware sinusoidal prior, VITA achieves state-of-the-art performance in predicting corn and soybean yields across 763 U.S. Corn Belt counties, particularly during extreme weather years.
## Citation
```bibtex
@inproceedings{hasan2026vita,
title={VITA: Variational Pretraining of Transformers for Climate-Robust Crop Yield Forecasting},
author={Adib Hasan and Mardavij Roozbehani and Munther Dahleh},
booktitle={Proceedings of the 40th AAAI Conference on Artificial Intelligence},
year={2026},
url={https://arxiv.org/abs/2508.03589},
}
```
|