Bridging the Gap between Requirements Engineering and Model Evaluation in Machine Learning

As the use of artificial intelligence (AI) systems in real-world settings has increased, so has demand for  assurances that AI-enabled systems perform as intended. Due to the  complexity of modern AI systems, the environments they are deployed in,  and the tasks they are designed to complete, providing such guarantees remains a challenge.  Defining and validating system behaviors through requirements engineering (RE) has been an integral component of software engineering since the 1970s. Despite the longevity of this  practice, requirements engineering for machine learning (ML) is not standardized and, as evidenced by interviews with ML practitioners and data scientists, is considered one of the hardest tasks in ML development. In this post, we define a simple evaluation framework centered around validating requirements and demonstrate this framework on an autonomous vehicle example.  We hope that this framework will serve as (1) a starting point for  practitioners to guide ML model development and (2) a touchpoint between  the software engineering and machine learning research communities.