DOI: 10.1177/14727978251361837 ISSN: 1472-7978

Curriculum engineering through AI: A neural architecture for aligning educational outcomes with labor market dynamics

Feifei Zhang

In today’s rapidly evolving global economy, the skills required by employers are constantly changing, placing increasing pressure on educational institutions to keep their curriculum relevant and responsive. Yet, academic programs comprising structured offerings like degrees, diplomas, and certificates frequently face challenges in adapting promptly to the changing demands of the labor market. This lag often leads to a sustained disconnect between the skills students acquire and those sought by employers. To address this challenge, this research introduces an innovative Artificial Intelligence (AI)-driven framework that leverages a Spatially-Aware Attention Long Short-Term Memory (SA-LSTM) architecture to harmonize academic curriculum with labor market trends. The framework processes two primary datasets: academic data, including curriculum outlines, course descriptions, student performance records, and labor market data, derived from online job postings, government employment reports, and industry white papers. The data undergoes a rigorous pre-processing pipeline, including cleaning, tokenization, and embedding using advanced Natural Language Processing (NLP) techniques. Term Frequency-Inverse Document Frequency (TF-IDF) is used to extract relevant features and feature-level fusion combines the features. The SA-LSTM model is the core of the system, incorporating spatial attention mechanisms to identify and model curriculum content and shifting skill requirements in the labor market. By leveraging the use of the attention mechanism, the model identifies contextual constraints in academic content and labor market trends to enable the identification of gaps in skills and new industry requirements. The SA-LSTM model generates curriculum recommendations from these results, suggesting new course modules, course modifications, and has better performance in accuracy with 96.2%, in comparison to baseline methods. This SA-LSTM offers a dynamic, data-driven solution to curriculum engineering, enabling schools to continuously adapt to the evolving needs of the job market.

More from our Archive