Optimizing Shortfall Risk Metric for Learning Regression Models Preprint link: https://arxiv.org/abs/2505.17777