Regularized adversarial examples for model interpretability

Yoel Shoshan, Vadim Ratner

As machine learning algorithms continue to improve, there is an increasing need for explaining why a model produces a certain prediction for a certain input. In recent years, sever…