The increasing demand for privacy-preserving, ethically aligned synthetic data generation in education has highlighted the limitations of existing tabular data generators. Traditional approaches often sacrifice fairness or privacy in pursuit of predictive accuracy, rendering them unsuitable for high-stakes academic settings. In this paper, we…

The main goal of learning analytics and early detection systems is to extract knowledge from student data to understand students' trends of activities towards success and risk and design intervention methods to improve learning performance and experience. However, many factors contribute to the challenge of designing and building effective…

This brief explores data science methods that workforce programs can use to predict participant success. With access to vast amounts of data on their programs, workforce training providers can leverage their management information systems (MIS) to understand and improve their programs' outcomes. By predicting which participants are at greater risk…