Abstract:

Machine learning algorithms are increasingly used in conjunction with optimization to guide decision making. A key challenge is aligning the machine learning loss with the decision-making loss. Existing solutions have limited flexibility and/or scale poorly to large datasets. We propose a principled decision-aware learning algorithm that uses a Taylor expansion of the optimal decision loss to derive the machine learning loss. Importantly, our approach only requires a simple re-weighting of the training data, allowing it to easily and scalably be incorporated into complex modern data science pipelines while producing sizable efficiency gains. We apply our framework to optimize the distribution of essential medicines in Sierra Leone in collaboration with their National Medical Supplies Agency. Out-of-sample results demonstrate that our end-to-end approach significantly reduces unmet demand across 1000+ health facilities throughout Sierra Leone.