A Support System to Help Teachers Design Course Plans Conforming to National Curriculum Guidelines

Yo Ehara

Abstract: This study addresses the challenge of ensuring the alignment of educational materials with the objectives of public education guidelines outlined in official documents. In Japan, “Courses of Study (COS),” issued by the government, provides a detailed outline for curricula in public and private education. Teachers ought to design lessons based on this outline. This can be a time-consuming and challenging task, particularly. Accordingly, this paper proposes a new support system for designing curricula that performs an automatic search of parts of COS corresponding to teaching materials. Computational retrieval via advanced deep transfer learning methods such as BERT and DeBERTa is expected to be effective and time-saving in the implementation of COS. Using a novel dataset linking textbook text with the COS, the proposed method annotates relevant sections of the COS. Experimental results showed that the proposed method could present the corresponding sections of the COS with a statistically significantly higher mean average precision (mAP) in both unsupervised and supervised learning compared to methods that do not consider semantic factors, such as edit distance.