MEGAN: Multi Explanation Graph Attention Network

Jonas Teufel, Luca Torresi, Patrick Nicholas Reiser, Pascal Friederich

Keywords: explainable artificial intelligence, interpretable machine learning, graph neural networks, attention network, graph regression, graph classification

TL;DR: Novel, self-explaining graph attention network features multiple explanation channels independent of task specifications to improve interpretability of graph regression and classification problems

Abstract: Explainable artificial intelligence (XAI) methods are expected to improve trust during human-AI interactions, provide tools for model analysis and extend human understanding of complex problems. Attention-based models are an important subclass of XAI methods, partly due to their full differentiability and the potential to improve explanations by means of explanation-supervised training. We propose the novel multi-explanation graph attention network (MEGAN). Our graph regression and classification model features multiple explanation channels, which can be chosen independently of the task specifications. We first validate our model on a synthetic graph regression dataset, where our model produces single-channel explanations with quality similar to GNNExplainer. Furthermore, we demonstrate the advantages of multi-channel explanations on one synthetic and two real-world datasets: The prediction of water solubility of molecular graphs and sentiment classification of movie reviews. We find that our model produces explanations consistent with human intuition, opening the way to learning from our model in less well-understood tasks.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning