Research Platform: Automated Reliability Decision Aid System (ARDAS)

Figure 1: ARDAS Platform

ARDAS is a human-in-the-loop predictive reliability ML-based simulation platform. ARDAS allows experimenters to obtain human performance data regarding Human and ML-based systems interaction and explore pertaining critical design questions on transparency (i.e., amount, content, and form). ARDAS will enable experimenters to configure the amount and content of information presented to the end-user, e.g., an experimenter can choose the logic that led to an automated prediction (including algorithm’s feature importance) to be displayed/not to be displayed to the end-user.

ARDAS was conceptually designed by David Quispe and Fahimeh Rajabiyazdi, and developed by David Quispe.

« Back to Research