A Design-to-Test Technique for Inclusion Coefficients in AI-Driven Systems
Yusuf U. Mshelia, Simon Tooswem Apeh, Charles Ikerionwu, Nachamada Vachaku Blamah, Azubuike I. Erike, Florence Elei · 2022 IEEE Nigeria 4th International Conference on Disruptive Technologies for Sustainable Development (NIGERCON) · 2022
AI drives the trending explosion in incorporated technologies like social interactive software systems. Such AI-Driven Systems (AIDS) in social technologies like Facebook, Google, Twitter, etc are unaided by human control and are rarely tested for articles of inclusion like identity, race, gender, privacy, ethics, etc. As a result, this paper proposed an architecture that designs AIDS to measure using Key Performance Index (KPI) metrics in testing for the coefficients of inclusion. As a research agenda, the conceptual architecture presented is designed in three layers; presentation layer, business logic layer and the data layer. These layers use in-built resources and/or API services (internal or external) for defining inclusion properties, article assortment and decision pattern for classifying articles or objects of inclusion.