Pareto-Optimal Learning-Augmented Algorithms for Online Conversion Problems

By incorporating predictions into design of OT A, we achieve the Pareto-optimal trade-off of consistency and robustness, i.e., no online algorithm can achieve a better consistency guarantee