Counterfactual Analysis is a powerful tool in Explainable Machine Learning. Given a classifier and a record, one seeks the smallest perturbation necessary to have the perturbed record, called the counterfactual explanation, classified in the desired class. Feature uncertainty in data reflects the inherent variability and noise present in real-world scenarios, and therefore, there is a … Read more
The paper studies distributionally robust optimization models with integer recourse. We develop a unified framework that provides finite tight convex relaxations under conic moment-based ambiguity sets and Wasserstein ambiguity sets. They provide tractable primal representations without relying on sampling or semi-infinite optimization. Beyond tractability, the relaxations offer interpretability that captures the criticality of recourse decisions. … Read more
Traditional end-to-end contextual robust optimization models are trained for specific contextual data, requiring complete retraining whenever new contextual information arrives. This limitation hampers their use in online decision-making problems such as energy scheduling, where multiperiod optimization must be solved every few minutes. In this paper, we propose a novel Data-Driven Contextual Uncertainty Set, which gives … Read more
Bulk-robust optimization is a recent paradigm for addressing problems in which the structure of a system is affected by uncertainty. It considers the case in which a finite and discrete set of possible failure scenarios is known in advance, and the decision maker aims to activate a subset of available resources of minimum cost so … Read more
Column liquid chromatography is an important technique applied in the production of biopharmaceuticals, specifically for the separation of biological macromolecules such as proteins. When setting up process conditions, it is crucial that the purity of the product is sufficiently high, even in the presence of perturbations in the process conditions, e.g., altered buffer salt concentrations. … Read more
Wasserstein distributionally robust optimization (DRO), a leading paradigm in data-driven decision-making, requires evaluating worst-case risk over a high-dimensional Wasserstein ball. We study when this worst-case evaluation admits an exact reduction to a one-dimensional formulation, in the sense that it can be carried out over a one-dimensional Wasserstein ball centered at the projected reference distribution. We … Read more
Contextual optimization enhances decision quality by leveraging side information to improve predictions of uncertain parameters. However, existing approaches face significant challenges when dealing with multimodal or mixtures of distributions. The inherent complexity of such structures often precludes an explicit functional relationship between the contextual information and the uncertain parameters, limiting the direct applicability of parametric … Read more
The Universal Classification (UC) problem seeks an optimal classifier from a universal policy space to minimize the expected 0-1 loss, also known as the misclassification risk. However, the conventional empirical risk minimization often leads to overfitting and poor out-of-sample performance. To address this limitation, we introduce the Distributionally Robust Universal Classification (DRUC) formulation, which incorporates … Read more
We introduce new algorithms to identify the trade-off (TRO) between adopting a distributional belief and hedging against ambiguity when modeling uncertainty in a capacitated fixed charge facility location problem (CFLP). We first formulate a TRO model for the CFLP (TRO-CFLP), which determines the number of facilities to open by minimizing the fixed establishment cost and … Read more
Logistic regression models are widely used in the social and behavioral sciences and in high-stakes domains, due to their simplicity and interpretability properties. At the same time, such domains are permeated by distribution shifts, where the distribution generating the data changes between training and deployment. In this paper, we study a distributionally robust logistic regression … Read more