Application of Causal Artificial Intelligence in Predicting the Effects of Continuous Improvement Measures

Authors: Lunliang Zhong

Abstract:

Continuous improvement is a crucial aspect of engineering education accreditation, vital for enhancing teaching quality and students' abilities. Typically, evaluating the effectiveness of continuous improvement measures requires a full academic year; Causal AI provides a distinct approach to address this issue. Using a self-assessment report from a particular university as a case study, we first constructed a course causal model, clarifying the causal relationships in teaching through the data generation process and structural equations. By treating continuous improvement measures as interventions, we can predict their impact on the student population. The counterfactual inference algorithm enables the prediction of these measures' effects on individual students. Both intervention inference and counterfactual inference indicate that implementing these continuous improvement measures can effectively improve students' final exam scores. The accuracy of these inferences was validated using student data collected before and after implementing the continuous improvement measures.

Keywords: engineering education professional certification, continuous improvement, causal ai, structural causal model