MIT researchers developed a method that helps a user to better understand a machine-learning model’s reasoning, and how that reasoning compares to that of a human. The technique enables a user to rapidly identify and analyze patterns in a model’s behavior.
Comments are closed.