Optimization and Interpretability of Graph Attention Networks for Small Sparse Graph Structures in Automotive Applications | IEEE Conference Publication | IEEE Xplore