WordDiffuser: Helping Beginners Cognize and Memorize Glyph-Semantic Pairs Using Diffusion

Yuhang Xu, Wanxu Xia, Yipeng Chen, Jiageng Liu, Weitao You

Abstract: This paper delves into symbolic cognition, emphasizing the evolution from ancient hieroglyphs to modern abstract characters. Acknowledging the cultural importance and learning challenges associated with these abstract characters, we propose WordDiffuser, an AI-aided framework generating images to help beginners understand the relationship between abstract characters and their meanings. Our research includes a review of the role of semantic association in education, an investigation identifying market shortcomings, and the creation of an interactive learning website. The effectiveness of our approach is substantiated through an experiment involving 22 participants, showing considerable improvements in abstract character recognition and memorization. This study paves the way for further advancements in technology-assisted learning of symbolic systems.1