Engendering Trust in Automated Feedback: A Two Step Comparison of Feedbacks in Gesture Based Learning

Advances in AI and Visual Recognition have paved the pathway for cutting edge research in Gesture Recognition. While automated feedback is able to open doors for newer opportunities in gesture based learning and practice, the effectiveness of these feedback as compared to manual feedback remains as a question in the minds of the users. For learners of American Sign Language (ASL), automated feedback generated by an application often causes a sense of apprehension because: a) learners are unaware of the automated feedback generation process, and b) learners fear that they can not trust the automated feedback as it may not be as good as the manual feedback. We use an ASL learning application that provides fine grained explainable feedback and follow a two step process to present a comparison between the automated feedback and the manual feedback provided by experts.