mlsanityData to 14 Oct 2026

Benchmarking and Enhancing Text-to-Image Models for Generating Visual Representations in Early Arithmetic Education

arXiv 2605.31212 ↗First posted 29 May 2026cs.CV cs.AI cs.CLcited by 0

Not on the map. Why?

Abstract (arXiv)

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.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 15 of its 73 references that are in the map. First posted on arXiv 29 May 2026.

Builds on

15 of 73 references.

References · 15

  1. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,181
  2. 2023Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detectioncited by 3,332
  3. 2025Emerging Properties in Unified Multimodal Pretrainingcited by 897
  4. 2024Measuring Multimodal Mathematical Reasoning with MATH-Vision Datasetcited by 526
  5. 2025Show-o2: Improved Native Unified Multimodal Modelscited by 250
  6. 2023Navigating Text-To-Image Customization: From LyCORIS Fine-Tuning to Model Evaluationcited by 109
  7. 2024CountGD: Multi-Modal Open-World Countingcited by 59
  8. 2023LLM Blueprint: Enabling Text-to-Image Generation with Complex and Detailed Promptscited by 54
  9. 2025How Well Does GPT-4o Understand Vision? Evaluating Multimodal Foundation Models on Standard Computer Vision Taskscited by 14
  10. 2025Text-to-Image Diffusion Models Cannot Count, and Prompt Refinement Cannot Helpcited by 12
  11. 2025Interactive Sketchpad: A Multimodal Tutoring System for Collaborative, Visual Problem-Solvingcited by 10
  12. 2023Enhancing Textbooks with Visuals from the Web for Improved Learningcited by 7
3 more
  1. 2024SIKeD: Self-guided Iterative Knowledge Distillation for mathematical reasoningcited by 6
  2. 2025Preliminary Explorations with GPT-4o(mni) Native Image Generationcited by 5
  3. 2025Can Vision-Language Models Solve Visual Math Equations?cited by 2

Built on by

Not cited yet.