Guardado en:
Detalles Bibliográficos
Autores principales: Kumar, Vishal, Mishra, Shubhra, Hao, Rebecca, Malik, Rizwaan, Broman, David, Demszky, Dorottya
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2511.08283
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866915611990294528
author Kumar, Vishal
Mishra, Shubhra
Hao, Rebecca
Malik, Rizwaan
Broman, David
Demszky, Dorottya
author_facet Kumar, Vishal
Mishra, Shubhra
Hao, Rebecca
Malik, Rizwaan
Broman, David
Demszky, Dorottya
contents Large Language Models (LLMs) are increasingly being adopted as tools for learning; however, most tools remain text-only, limiting their usefulness for domains where visualizations are essential, such as mathematics. Recent work shows that LLMs are capable of generating code that compiles to educational figures, but a major bottleneck remains: scalable evaluation of these diagrams. We address this by proposing DiagramIR: an automatic and scalable evaluation pipeline for geometric figures. Our method relies on intermediate representations (IRs) of LaTeX TikZ code. We compare our pipeline to other evaluation baselines such as LLM-as-a-Judge, showing that our approach has higher agreement with human raters. This evaluation approach also enables smaller models like GPT-4.1-Mini to perform comparably to larger models such as GPT-5 at a 10x lower inference cost, which is important for deploying accessible and scalable education technologies.
format Preprint
id arxiv_https___arxiv_org_abs_2511_08283
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle DiagramIR: An Automatic Pipeline for Educational Math Diagram Evaluation
Kumar, Vishal
Mishra, Shubhra
Hao, Rebecca
Malik, Rizwaan
Broman, David
Demszky, Dorottya
Artificial Intelligence
Large Language Models (LLMs) are increasingly being adopted as tools for learning; however, most tools remain text-only, limiting their usefulness for domains where visualizations are essential, such as mathematics. Recent work shows that LLMs are capable of generating code that compiles to educational figures, but a major bottleneck remains: scalable evaluation of these diagrams. We address this by proposing DiagramIR: an automatic and scalable evaluation pipeline for geometric figures. Our method relies on intermediate representations (IRs) of LaTeX TikZ code. We compare our pipeline to other evaluation baselines such as LLM-as-a-Judge, showing that our approach has higher agreement with human raters. This evaluation approach also enables smaller models like GPT-4.1-Mini to perform comparably to larger models such as GPT-5 at a 10x lower inference cost, which is important for deploying accessible and scalable education technologies.
title DiagramIR: An Automatic Pipeline for Educational Math Diagram Evaluation
topic Artificial Intelligence
url https://arxiv.org/abs/2511.08283