Robust explainability: a tutorial on gradient-based attribution methods for deep neural networks

DOI：10.48550/arxiv.2107.11400