Math2Visual: A Framework for Generating Pedagogically Meaningful Visuals for Teaching Math Word Problems

ACL ARR 2025 February Submission720 Authors

Abstract: Visuals are valuable tools in teaching math word problems (MWPs), helping young learners interpret textual descriptions into mathematical expressions before solving them. However, creating such visuals is labor-intensive and we lack automated methods. In this paper, we present Math2Visual, an automatic framework for generating pedagogically meaningful visuals from MWP text descriptions. Math2Visual leverages a pre-defined visual language and a structured design space for visuals, informed by math teachers, to effectively capture the essential mathematical relationships within MWPs. Using Math2Visual, we construct an annotated dataset of 1,903 visuals and evaluate Text-to-Image (TTI) models in generating visuals that align with our design. We further fine-tune various TTI models with our dataset, demonstrating improvements in educational visual generation. Our work establishes a new benchmark for automated pedagogical meaningful visual generation and offers insights into the challenge of generating multimodal educational content.

Paper Type: Long

Research Area: NLP Applications

Research Area Keywords: educational applications, multimodal applications, human-centered evaluation

Contribution Types: Model analysis & interpretability, Publicly available software and/or pre-trained models, Data resources

Languages Studied: English

Submission Number: 720