<ul><li>Spatiotemporal Graph Learning (SGL) under Zero-Inflated Distribution (ZID) is crucial for urban risk management tasks.</li><li>Traditional adversarial training (AT) exacerbates performance disparities between majority and minority classes under ZID.</li><li>The proposed MinGRE framework addresses the performance disparity by reweighting gradients and enhancing representations of the minority class.</li><li>MinGRE achieves enhanced robustness and reduces the performance disparity across classes compared to existing baselines.</li></ul>

Alleviating Performance Disparity in Adversarial Spatiotemporal Graph Learning Under Zero-Inflated Distribution

Discover more