Recent work on training neural networks to have robust attributions, to improve their trustworthiness and resilience to adversarial attacks.…