GREASE: Generate Factual and Counterfactual Explanations for GNN-based Recommendations

TL;DR

GREASE addresses the challenge of explaining GNN-based recommendations by producing both factual (sufficient) and counterfactual (necessary) explanations for user-item rankings. It operates in a black-box setting by learning a local surrogate GNN on the $l$-hop neighborhood around a target pair and optimizing adjacency perturbations to extract concise explanations via $A^{FA}$ and $A^{CF}$. The method is validated on LastFM and Yelp with LightGCN and NGCF, showing improvements over baselines in PS, PN, and EC, indicating more informative and compact explanations. This work lays a foundation for ranking-oriented explanations in graph-based recommenders and suggests avenues for jointly integrating factual and counterfactual explanations in future work.

Abstract

Recently, graph neural networks (GNNs) have been widely used to develop successful recommender systems. Although powerful, it is very difficult for a GNN-based recommender system to attach tangible explanations of why a specific item ends up in the list of suggestions for a given user. Indeed, explaining GNN-based recommendations is unique, and existing GNN explanation methods are inappropriate for two reasons. First, traditional GNN explanation methods are designed for node, edge, or graph classification tasks rather than ranking, as in recommender systems. Second, standard machine learning explanations are usually intended to support skilled decision-makers. Instead, recommendations are designed for any end-user, and thus their explanations should be provided in user-understandable ways. In this work, we propose GREASE, a novel method for explaining the suggestions provided by any black-box GNN-based recommender system. Specifically, GREASE first trains a surrogate model on a target user-item pair and its $l$-hop neighborhood. Then, it generates both factual and counterfactual explanations by finding optimal adjacency matrix perturbations to capture the sufficient and necessary conditions for an item to be recommended, respectively. Experimental results conducted on real-world datasets demonstrate that GREASE can generate concise and effective explanations for popular GNN-based recommender models.