Landscape of Requirements Engineering for Machine Learning-based AI Systems

Techniques and practices in RE are not well researched, although problems and the research challenges on requirements engineering (RE) for machine learning-based systems (MLS) are evaluated via empirical case studies. A systematic literature review of RE for MLS was conducted to guide practitioners and researchers to design and research effective RE for ML systems and software. We identified 32 papers. Although many studies have been recently conducted, problem statements and research challenges remain. Future studies should include the monitoring requirements for concept drifts and how domain experts collaborate with ML experts and engineers.