Improving Student Learning Outcome Tracing at HBCUs Using Tabular Generative AI and Deep Knowledge Tracing

Historically Black Colleges and Universities in the United States serve a vital role in providing educational opportunities and training, particularly for underrepresented students, facing a challenge of lower retention and graduation rates compared to other institutions. To overcome this challenge, this study explores the application of generative artificial intelligence models to generate synthetic data, augmenting real datasets to improve student learning outcome tracing at these colleges and universities using Deep Knowledge Tracing techniques, which potentially offers actionable insights to identify at-risk students and enables proactive interventions to enhance retention and graduation rates in Science, Technology, Engineering and Math education. Utilizing two years of educational data from Prairie View A&M University, it applied data augmentation with tabular generative artificial intelligence models. The experimental results indicate that augmenting training data with synthetic samples generated by these models improved tracing performance measured by AUC and accuracy by approximately 5% and 3%, respectively, underscoring the potential of synthetic data to enhance the monitoring of student learning outcomes in diverse educational contexts. These findings highlight the critical role of data augmentation through generative artificial intelligence in improving the student learning outcome tracing, offering valuable insights for strategies to enhance retention and graduation rates.