CME-CAD: Heterogeneous Collaborative Multi-Expert Reinforcement Learning for CAD Code Generation

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Niu, Ke, Yu, Haiyang, Chen, Zhuofan, Yao, Zhengtao, Jia, Weitao, Ge, Xiaodong, Tang, Jingqun, Cui, Benlei, Li, Bin, Xue, Xiangyang
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866918265315393536
author Niu, Ke
Yu, Haiyang
Chen, Zhuofan
Yao, Zhengtao
Jia, Weitao
Ge, Xiaodong
Tang, Jingqun
Cui, Benlei
Li, Bin
Xue, Xiangyang
author_facet Niu, Ke
Yu, Haiyang
Chen, Zhuofan
Yao, Zhengtao
Jia, Weitao
Ge, Xiaodong
Tang, Jingqun
Cui, Benlei
Li, Bin
Xue, Xiangyang
contents Computer-Aided Design (CAD) is essential in industrial design, but the complexity of traditional CAD modeling and workflows presents significant challenges for automating the generation of high-precision, editable CAD models. Existing methods that reconstruct 3D models from sketches often produce non-editable and approximate models that fall short of meeting the stringent requirements for precision and editability in industrial design. Moreover, the reliance on text or image-based inputs often requires significant manual annotation, limiting their scalability and applicability in industrial settings. To overcome these challenges, we propose the Heterogeneous Collaborative Multi-Expert Reinforcement Learning (CME-CAD) paradigm, a novel training paradigm for CAD code generation. Our approach integrates the complementary strengths of these models, facilitating collaborative learning and improving the model's ability to generate accurate, constraint-compatible, and fully editable CAD models. We introduce a two-stage training process: Multi-Expert Fine-Tuning (MEFT), and Multi-Expert Reinforcement Learning (MERL). Additionally, we present CADExpert, an open-source benchmark consisting of 17,299 instances, including orthographic projections with precise dimension annotations, expert-generated Chain-of-Thought (CoT) processes, executable CADQuery code, and rendered 3D models.
format Preprint
id arxiv_https___arxiv_org_abs_2512_23333
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle CME-CAD: Heterogeneous Collaborative Multi-Expert Reinforcement Learning for CAD Code Generation
Niu, Ke
Yu, Haiyang
Chen, Zhuofan
Yao, Zhengtao
Jia, Weitao
Ge, Xiaodong
Tang, Jingqun
Cui, Benlei
Li, Bin
Xue, Xiangyang
Computer Vision and Pattern Recognition
Computer-Aided Design (CAD) is essential in industrial design, but the complexity of traditional CAD modeling and workflows presents significant challenges for automating the generation of high-precision, editable CAD models. Existing methods that reconstruct 3D models from sketches often produce non-editable and approximate models that fall short of meeting the stringent requirements for precision and editability in industrial design. Moreover, the reliance on text or image-based inputs often requires significant manual annotation, limiting their scalability and applicability in industrial settings. To overcome these challenges, we propose the Heterogeneous Collaborative Multi-Expert Reinforcement Learning (CME-CAD) paradigm, a novel training paradigm for CAD code generation. Our approach integrates the complementary strengths of these models, facilitating collaborative learning and improving the model's ability to generate accurate, constraint-compatible, and fully editable CAD models. We introduce a two-stage training process: Multi-Expert Fine-Tuning (MEFT), and Multi-Expert Reinforcement Learning (MERL). Additionally, we present CADExpert, an open-source benchmark consisting of 17,299 instances, including orthographic projections with precise dimension annotations, expert-generated Chain-of-Thought (CoT) processes, executable CADQuery code, and rendered 3D models.
title CME-CAD: Heterogeneous Collaborative Multi-Expert Reinforcement Learning for CAD Code Generation
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2512.23333