Uncertainty-Guided Label Rebalancing for CPS Safety Monitoring

Researchers have introduced U-Balance, a method for improving safety monitoring in Cyber-Physical Systems by addressing class imbalance through behavioral uncertainty. This technique trains an uncertainty predictor to probabilistically relabel safe instances as unsafe when they exhibit high uncertainty, enhancing the minority class without creating synthetic data. Developers can leverage this approach to enhance predictive models' accuracy and efficiency in scenarios with rare but critical events.

Read the full article at arXiv cs.LG (ML)

---

Want to create content about this topic? Use Nemati AI tools to generate articles, social posts, and more.