DeciCAD improves CAD actions with harness-guided learning policies
DeciCAD: Harness-Grounded Learning of Action Policies for CAD Agents
Artificial IntelligenceComputation and LanguageMachine LearningRobotics
Summary
Computer-aided design (CAD) involves complex and detailed steps to create and modify designs. The authors studied how defining the possible actions and decision space (called a harness) affects how AI agents perform and learn in CAD tasks. They created DeciCAD, a special harness that closely matches CAD operations and lets AI learn more effectively through examples and feedback. This approach led to better accuracy and fewer unnecessary outputs than general coding methods. Their work shows that carefully choosing the decision framework can improve AI performance in professional design software.
What this means in practice
- •For cad software developers: Enhance AI assistants in CAD tools by integrating domain-specific action harnesses to achieve more accurate design automation and reduce unnecessary command outputs.
- •For robotics control engineers: Use domain-tailored harnesses to improve action policy learning in robots performing long-horizon, constrained tasks similar to CAD workflows.
Authors
Yufeng Peng, Yixin Zhang, Ziyi Liu, Zhikun Xu, Jiarong Yang, Sijie Wang, Wenjie Feng, Qiyu Kang
Abstract
Large Language Model agents operate through closed-loop action policies whose decision space is defined by an agent harness. We investigate how this operating space shapes agent behavior and policy learning in computer-aided design (CAD), a professional domain characterized by persistent state, geometric constraints, and long-horizon procedures. We instantiate this perspective in DeciCAD, a domain-native harness that exposes semantic CAD actions, decision-oriented observations, environment feedback, and persistent state transitions. Building on this harness, we develop a learning pipeline that extracts causal decision supervision from interaction trajectories and improves the policy through supervised fine-tuning (SFT) and geometry-grounded online reinforcement learning (RL). Across four CAD benchmarks, DeciCAD achieves the best F1 and intersection over union (IoU) while reducing model-output tokens by 45--88% relative to a general-purpose coding harness. Under matched supervision, its SFT policy improves F1 from 0.6883 to 0.7437 and outperforms the coding policy; subsequent online RL further raises geometry availability from 66.5% to 73.5% and F1 to 0.7880. These results demonstrate that the decision space induced by a harness materially affects both agent execution and policy learning in CAD.