Professor Dolores Romero Morales and co-authors win an award
Dolores Romero Morales (professor of Operations Research at the Copenhagen Business School), Emilio Carrizosa (professor of Statistics and Operations Research at the University of Seville) and Jasone Ramírez-Ayerbe (predoctoral researcher in Statistics and Operations Research at the University of Seville) won an award for Best contribution in Statistics and Operations Research applied to Data Science and Big Data for their paper Mathematical optimization modelling for group counterfactual explanations, published in the European Journal of Operational Research.
This is an original contribution to the field of Explainable Artificial Intelligence (XAI) for classification using the important new idea of Counterfactual Analysis. In the simplest case of understanding a classification method for some records into two given classes, a useful procedure is to associate to each record a so-called counterfactual explanation, that is, a small change in the record that makes it to have a high probability of being classified in the opposite class. In this article a collection of new models with combinatorial structure are presented for situations in which the same counterfactual solution can be used for several records, a problem that has not been studied. These new problems can be easily addressed with Mixed Integer (Nonlinear) Optimization solvers and the article analyses new methods using Mathematical Optimization ideas to solve them.