Explaining Preferences with Shapley Values

While preference modelling is becoming one of the pillars of machine learning, the problem of preference explanation remains challenging and underexplored. In this paper, we propose \textsc{Pref-SHAP}, a Shapley value-based model explanation framework for pairwise comparison data. We derive the appropriate value functions for preference models and further extend the framework to model and explain \emph{context specific} information, such as the surface type in a tennis game. To demonstrate the utility of \textsc{Pref-SHAP}, we apply our method to a variety of synthetic and real-world datasets and show that richer and more insightful explanations can be obtained over the baseline.

Authors

• Robert Hu
• Siu Lun Chau CISPA Helmholtz Center for Information Security, Saarbrücken, Germany
• Jaime Ferrando Huertas
• Dino Sejdinovic University of Adelaide

Keywords

No keywords are indexed for this paper.