Architecting for Fairness
Urjaswala Vora · 2025
Fairness in a software system has majorly been analyzed to validate or mitigate biases and to reduce, rather ideally to remove, unfair program behaviors. A lot of research work has been covered in learning-based systems too but ensuring the mitigation of fairness concerns at the system analysis and design level for every system is neither studied nor practiced. We propose the analysis and design model for fairness to be included in architecture design phase of software development lifecycle models which in turn leads to design of metrics for measuring the fairness as a software quality and subsequently the fairness debt in a software architecture.