Conformal Contextual Robust Optimization
This addresses uncertainty misspecification in safety-critical decision-making, offering a less conservative and more interpretable approach, though it appears incremental by building on existing conformal and generative methods.
The paper tackles the problem of overly conservative uncertainty regions in predict-then-optimize decision-making, proposing a framework that uses conformal prediction with generative models to achieve distribution-free coverage and provide visual explanations for optimal decisions, demonstrated on benchmark tasks and a vehicle routing application.
Data-driven approaches to predict-then-optimize decision-making problems seek to mitigate the risk of uncertainty region misspecification in safety-critical settings. Current approaches, however, suffer from considering overly conservative uncertainty regions, often resulting in suboptimal decisionmaking. To this end, we propose Conformal-Predict-Then-Optimize (CPO), a framework for leveraging highly informative, nonconvex conformal prediction regions over high-dimensional spaces based on conditional generative models, which have the desired distribution-free coverage guarantees. Despite guaranteeing robustness, such black-box optimization procedures alone inspire little confidence owing to the lack of explanation of why a particular decision was found to be optimal. We, therefore, augment CPO to additionally provide semantically meaningful visual summaries of the uncertainty regions to give qualitative intuition for the optimal decision. We highlight the CPO framework by demonstrating results on a suite of simulation-based inference benchmark tasks and a vehicle routing task based on probabilistic weather prediction.