Conformal Load Prediction with Transductive Graph Autoencoders
Abstract
Predicting edge weights on graphs has various applications, from transportation systems to social networks. This paper describes a Graph Neural Network (GNN) approach for edge weight prediction with guaranteed coverage. We leverage conformal prediction to calibrate the GNN outputs and produce valid prediction intervals. We handle data heteroscedasticity through error reweighting and Conformalized Quantile Regression (CQR). We compare the performance of our method against baseline techniques on real-world transportation datasets. Our approach has better coverage and efficiency than all baselines and showcases robustness and adaptability.
Cite
Text
Luo and Colombo. "Conformal Load Prediction with Transductive Graph Autoencoders." Machine Learning, 2025. doi:10.1007/S10994-024-06713-WMarkdown
[Luo and Colombo. "Conformal Load Prediction with Transductive Graph Autoencoders." Machine Learning, 2025.](https://mlanthology.org/mlj/2025/luo2025mlj-conformal/) doi:10.1007/S10994-024-06713-WBibTeX
@article{luo2025mlj-conformal,
title = {{Conformal Load Prediction with Transductive Graph Autoencoders}},
author = {Luo, Rui and Colombo, Nicolò},
journal = {Machine Learning},
year = {2025},
pages = {54},
doi = {10.1007/S10994-024-06713-W},
volume = {114},
url = {https://mlanthology.org/mlj/2025/luo2025mlj-conformal/}
}