Self-explainable deep neural networks are a recent class of models that can output ante-hoc local explanations that are faithful to the model's reasoning, and as such represent a step forward toward filling the gap between expressiveness and interpretability. Self-explainable graph neural networks (GNNs) aim at achieving the same in the context of graph data. This begs the question: do these models fulfill their implicit guarantees in terms of faithfulness? In this extended abstract, we analyze the faithfulness of several self-explainable GNNs using different measures of faithfulness, identify several limitations -- both in the models themselves and in the evaluation metrics -- and outline possible ways forward.
How Faithful are Self-Explainable GNNs? / Marc, Christiansen; Villadsen, Lea; Zhong, Zhiqiang; Teso, Stefano; Mottin, Davide. - ELETTRONICO. - (2023). (Intervento presentato al convegno LOG23 tenutosi a Multiple locations nel 27 November 2023 - 30 November 2023).
How Faithful are Self-Explainable GNNs?
Teso, Stefano;Mottin, DavideUltimo
2023-01-01
Abstract
Self-explainable deep neural networks are a recent class of models that can output ante-hoc local explanations that are faithful to the model's reasoning, and as such represent a step forward toward filling the gap between expressiveness and interpretability. Self-explainable graph neural networks (GNNs) aim at achieving the same in the context of graph data. This begs the question: do these models fulfill their implicit guarantees in terms of faithfulness? In this extended abstract, we analyze the faithfulness of several self-explainable GNNs using different measures of faithfulness, identify several limitations -- both in the models themselves and in the evaluation metrics -- and outline possible ways forward.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione