Research map / Large language models / Language, safety and interpretability
Pointer-CAD: Unifying B-Rep and Command Sequences via Pointer-based Edges & Faces Selection
Dacheng Qi, Chenyu Wang, Jingwei Xu, Tianzhe Chu, Zibo Zhao, Wen Liu, Wenrui Ding, Yi Ma, Shenghua Gao
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 B-rep GNN encoder instead of 3-layer MLP (ablation) | Pointer-CAD-0.5B · Recap-DeepCAD | Arc F1: 67.14 -> 85.70 | Improved |
| add B-rep GNN encoder instead of 3-layer MLP (ablation) | Pointer-CAD-0.5B · Recap-DeepCAD | IR (invalid ratio): 22.73 -> 15.02 | Improved |
| add GNN B-rep encoder with multi-step generation (ablation) | Text2CAD · Recap-DeepCAD | CD mean: 17.48 -> 14.33 | Improved |
| use pointer-based selection and multi-step B-rep conditioning vs Text2CAD | Text2CAD · Recap-DeepCAD | SegE: 0.44 -> 0.13 | Improved |
| scale Pointer-CAD from 0.5B to 1.5B | Pointer-CAD · Recap-DeepCAD | SegE: 0.13 -> 0.11 | No clear effect |
| scale Pointer-CAD from 0.5B to 1.5B | Pointer-CAD · Recap-OmniCAD+ | Fillet F1: 82.54 -> 89.85 | Improved |
| pointer-based CAD vs CADmium-7B (larger LLM) | CADmium-7B · Recap-DeepCAD | CD mean: 10.53 -> 3.81 | Improved |
| use Pointer-CAD instead of prompting general LLMs for CadQuery | Gemini 2.5 Pro · Recap-DeepCAD-2K | IR (invalid ratio): 24.95 -> 14.79 | Improved |
| use command sequence with pointers vs CadQuery code representation | Text-to-CadQuery · generation time | time: 6.14 -> 2.13 | Improved |
| pointer-based edge selection enabling chamfer and fillet | Pointer-CAD-0.5B · Recap-OmniCAD+ | Chamfer F1 | Improved |
| use Pointer-CAD vs Text2CAD on arc F1 | Text2CAD · Recap-DeepCAD | Arc F1: 45.19 -> 85.70 | Improved |
Closest papers by meaning
- CAD-Llama: Leveraging Large Language Models for Computer-Aided Design Parametric 3D Model Generation (CVPR 2025 · similarity 0.91)
- CAD-Coder: Text-to-CAD Generation with Chain-of-Thought and Geometric Reward (NeurIPS 2025 · similarity 0.89)
- FlexCAD: Unified and Versatile Controllable CAD Generation with Fine-tuned Large Language Models (ICLR 2025 · similarity 0.88)
- CAD-Editor: A Locate-then-Infill Framework with Automated Training Data Synthesis for Text-Based CAD Editing (ICML 2025 · similarity 0.88)
- CAD-GPT: Synthesising CAD Construction Sequence with Spatial Reasoning-Enhanced Multimodal LLMs (AAAI 2025 · similarity 0.88)
- From 2D CAD Drawings to 3D Parametric Models: A Vision-Language Approach (AAAI 2025 · similarity 0.87)
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.
