Applying AI Tools to Develop a Curriculum Based on Expected Learning Outcomes and Personalize Learning Program for Students at the University of Languages and International Studies

The higher education system in Vietnam is undergoing a significant shift from training based on university capacity to training based on labor market demands. In a developing economy dominated by small and medium-sized enterprises, it is a big challenge to train graduates to meet changing and very diverse competence requirements. AI and machine learning tools are applied in three stages: (a) processing survey data: Expected learning outcomes (pELOs) are quantified into credit values, with each module's contribution determined using the apriori algorithm and expert methods; (b) Optimizing framework curricula (FC): A genetic algorithm identifies module combinations that meet all pELOs while minimizing redundancy within a specified study duration; (c) Framework curriculum adjustment (FCA): An FCA tool, using genetic algorithms, enables schools to update FC annually and allows learners to personalize their programs. WEKA is used to implement the apriori algorithm (https://www.cs.waikato.ac.nz/ml/weka). The PASCAL language is used to write GA and its associated subroutines. Foreign language bachelor's degree programs at the University of Languages and International Studies Vietnam National University, Hanoi, (ULIS-VNU) were used to test algorithms and procedures. According to the calculations, present FCs have caused overstudy and can be modified for every employment post to reduce the surplus credit values that have accrued. Furthermore, FCA can assist in making the curriculum more flexible so that students can more easily switch out FC modules based on their skills and circumstances while still meeting all of the stated ELOs. Under project number N.21.13, this research piece was finished with assistance from ULIS-VNU.