Task-based End-to-end Model Learning in Stochastic Optimization
This addresses the problem of suboptimal model performance in real-world applications like energy and inventory management, though it is incremental as it builds on existing stochastic programming frameworks.
The paper tackles the mismatch between training criteria and ultimate task objectives in machine learning models used within larger processes by proposing an end-to-end approach that directly optimizes for task-based objectives in stochastic programming. The approach outperforms traditional modeling and black-box policy optimization in inventory stock, electrical grid scheduling, and energy storage arbitrage tasks.
With the increasing popularity of machine learning techniques, it has become common to see prediction algorithms operating within some larger process. However, the criteria by which we train these algorithms often differ from the ultimate criteria on which we evaluate them. This paper proposes an end-to-end approach for learning probabilistic machine learning models in a manner that directly captures the ultimate task-based objective for which they will be used, within the context of stochastic programming. We present three experimental evaluations of the proposed approach: a classical inventory stock problem, a real-world electrical grid scheduling task, and a real-world energy storage arbitrage task. We show that the proposed approach can outperform both traditional modeling and purely black-box policy optimization approaches in these applications.