Elayne Ruane

Assessing and comparing the quality of conversational agents, e.g. chatbots, before roll-out to users is a difficult task. This work aims to develop an automated divergent testing framework to assess the robustness of an agent or agents using a black-box approach.