Research map / Large language models / Agents, code and math
GeoLoom: High-quality Geometric Diagram Generation from Textual Input
Xiaojing Wei, Ting Zhang, Wei He, Jingdong Wang, Hua Huang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add validation-based filters to autoformalization (ablation) | GeoLoom training-free (DeepSeek-V3) · GeoNF test | LCI/ADI average: 1.602 -> 1.337 | Improved |
| add validation-based filters (ablation) | GeoLoom training-free · GeoNF test | LCI: 1.340 -> 0.962 | Improved |
| add validation-based filters (ablation) | GeoLoom training-free · GeoNF test | generation time: 15.945 -> 13.379 | Improved |
| use only length constraints instead of full constraints (ablation) | GeoLoom coordinate solver · GeoNF test | LCI: 0.870 -> 0.651 | Mixed |
| use only length constraints instead of full constraints (ablation) | GeoLoom coordinate solver · GeoNF test | time: 15.782 -> 2.844 | Improved |
| use only angle constraints instead of full constraints (ablation) | GeoLoom coordinate solver · GeoNF test | ADI: 1.537 -> 1.577 | Hurt |
| use fine-tuned Qwen2.5-7b instead of Qwen2.5-14b | GeoLoom autoformalization · GeoNF test | manual accuracy (%): 83.72 -> 85.34 | Improved |
| fine-tune Qwen2.5-7b instead of training-free DeepSeek-V3 | GeoLoom autoformalization · GeoNF test | manual accuracy (%): 81.16 -> 85.34 | Improved |
| use LLaMa3.1-8b instead of LLaMa3-8b | GeoLoom fine-tuned · GeoNF test | manual accuracy (%): 75.58 -> 79.06 | Improved |
| use GeoLoom fine-tuned instead of training-free | GeoLoom · User study | quality preference (%): 39.2 -> 55.1 | Improved |
| use GeoLoom instead of Seedream3.0 or AutomaTikZ | text-to-diagram · User study | alignment preference (%): 0.2 -> 53.9 | Improved |
| increase inner and outer iterations Q and T (ablation) | Monte Carlo solver · GeoNF test | quality and run time | Mixed |
Closest papers by meaning
- Geoint-R1: Formalizing Multimodal Geometric Reasoning with Dynamic Auxiliary Constructions (CVPR 2026 · similarity 0.89)
- Synthesizing Multimodal Geometry Datasets from Scratch and Enabling Visual Alignment via Plotting Code (ICML 2026 · similarity 0.87)
- GeoX: Geometric Problem Solving Through Unified Formalized Vision-Language Pre-training (ICLR 2025 · similarity 0.87)
- Do Large Language Models Truly Understand Geometric Structures? (ICLR 2025 · similarity 0.87)
- Hilbert-Geo: Solving Solid Geometric Problems by Neural-Symbolic Reasoning (CVPR 2026 · similarity 0.86)
- GeoTikzBridge: Advancing Multimodal Code Generation for Geometric Perception and Reasoning (CVPR 2026 · similarity 0.86)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
