Benchmarking and Enhancing Text-to-Image Models for Generating Visual Representations in Early Arithmetic Education
This work addresses the problem of generating accurate and pedagogically sound visual representations for early arithmetic education, which is crucial for teachers and content creators.
This paper introduces equation-to-visual generation, a new task requiring AI models to create pedagogically meaningful visuals from arithmetic equations while preserving numerical and relational structures. They found that current text-to-image models frequently fail, primarily due to incorrect object counts and broken relational structures, but enhancement strategies improved performance.
AI systems are increasingly used to support educational content creation, yet it remains unclear whether they can generate outputs that faithfully represent the pedagogical concepts they are intended to teach. Thus, we introduce equation-to-visual generation, a task that, in contrast to conventional image generation, requires producing pedagogically meaningful visuals from arithmetic equations while precisely preserving their numerical and relational structure. Informed by interviews with teachers and an analysis of educational materials, we construct E2V-Bench, a benchmark spanning four pedagogically grounded visual types, along with automatic metrics for evaluating visual correctness. Our evaluation reveals that recent text-to-image (T2I) models frequently fail on this task, with errors dominated by incorrect object counts and broken relational structure. Building on this, we explore benchmark-guided enhancement strategies. These strategies improve representative models, while the remaining gap calls for stronger numerical and relational grounding in future T2I models.