Training for Stable Explanation for Free Chao Chen 1 Rufeng Chen 3 Guixiang Ma4

Neural Information Processing Systems 

To foster trust in machine learning models, explanations must be faithful and stable for consistent insights.