Decision-Theoretic Approaches for Improved Learning-Augmented Algorithms
This work addresses the challenge of systematically evaluating learning-augmented algorithms for researchers and practitioners in online decision-making, though it appears incremental by building on existing prediction-based methods.
The paper tackles the problem of designing and analyzing algorithms that incorporate machine-learned predictions by introducing decision-theoretic metrics to quantify performance across prediction errors, applying this framework to ski-rental, one-max search, and contract scheduling problems.
We initiate the systematic study of decision-theoretic metrics in the design and analysis of algorithms with machine-learned predictions. We introduce approaches based on both deterministic measures such as distance-based evaluation, that help us quantify how close the algorithm is to an ideal solution, and stochastic measures that balance the trade-off between the algorithm's performance and the risk associated with the imperfect oracle. These approaches allow us to quantify the algorithm's performance across the full spectrum of the prediction error, and thus choose the best algorithm within an entire class of otherwise incomparable ones. We apply our framework to three well-known problems from online decision making, namely ski-rental, one-max search, and contract scheduling.