Optimization and Interpretability of Graph Attention Networks for Small Sparse Graph Structures in Automotive Applications
This work addresses incremental improvements for automotive systems using graph neural networks, focusing on better parameter learning and interpretability.
The paper tackled optimization issues in Graph Attention Networks (GATv2) for small sparse graphs in automotive applications, proposing architectural modifications that improved prediction performance and robustness in node-level regression tasks.
For automotive applications, the Graph Attention Network (GAT) is a prominently used architecture to include relational information of a traffic scenario during feature embedding. As shown in this work, however, one of the most popular GAT realizations, namely GATv2, has potential pitfalls that hinder an optimal parameter learning. Especially for small and sparse graph structures a proper optimization is problematic. To surpass limitations, this work proposes architectural modifications of GATv2. In controlled experiments, it is shown that the proposed model adaptions improve prediction performance in a node-level regression task and make it more robust to parameter initialization. This work aims for a better understanding of the attention mechanism and analyzes its interpretability of identifying causal importance.