Draw it like Euclid: Teaching transformer models to generate CAD profiles using ruler and compass construction steps
This addresses the challenge of precise and editable CAD profile generation for designers, offering a novel approach that enables parametric editing with floating-point precision, though it is incremental in applying transformer and reinforcement learning techniques to a specific domain.
The paper tackles the problem of generating Computer Aided Design (CAD) profiles by using transformer models to produce sequences of geometric construction steps, such as curve offsetting and rotations, starting from designer input. The result shows that adding these construction steps improves generation quality, similar to chain-of-thought in language models, and reinforcement learning further enhances performance across multiple metrics.
We introduce a new method of generating Computer Aided Design (CAD) profiles via a sequence of simple geometric constructions including curve offsetting, rotations and intersections. These sequences start with geometry provided by a designer and build up the points and curves of the final profile step by step. We demonstrate that adding construction steps between the designer's input geometry and the final profile improves generation quality in a similar way to the introduction of a chain of thought in language models. Similar to the constraints in a parametric CAD model, the construction sequences reduce the degrees of freedom in the modeled shape to a small set of parameter values which can be adjusted by the designer, allowing parametric editing with the constructed geometry evaluated to floating point precision. In addition we show that applying reinforcement learning to the construction sequences gives further improvements over a wide range of metrics, including some which were not explicitly optimized.