The Importance of Model Inspection for Better Understanding Performance Characteristics of Graph Neural Networks
It addresses the need for better model inspection tools for practitioners in medical imaging, though it is incremental as it applies existing inspection methods to a specific domain.
This study tackled the problem of understanding performance characteristics in graph neural networks for brain shape classification, finding that test accuracy alone is insufficient to identify biases and non-discriminative features, with substantial differences in feature embeddings observed across model layers.
This study highlights the importance of conducting comprehensive model inspection as part of comparative performance analyses. Here, we investigate the effect of modelling choices on the feature learning characteristics of graph neural networks applied to a brain shape classification task. Specifically, we analyse the effect of using parameter-efficient, shared graph convolutional submodels compared to structure-specific, non-shared submodels. Further, we assess the effect of mesh registration as part of the data harmonisation pipeline. We find substantial differences in the feature embeddings at different layers of the models. Our results highlight that test accuracy alone is insufficient to identify important model characteristics such as encoded biases related to data source or potentially non-discriminative features learned in submodels. Our model inspection framework offers a valuable tool for practitioners to better understand performance characteristics of deep learning models in medical imaging.