SketchDNN: Joint Continuous-Discrete Diffusion for CAD Sketch Generation

Sathvik Reddy Chereddy, John Femiani

TL;DR: We present SketchDNN, a generative model for CAD sketches and introduce Gaussian-Softmax diffusion.

Abstract: We present SketchDNN, a generative model for synthesizing CAD sketches that jointly models both continuous parameters and discrete class labels through a unified continuous-discrete diffusion process. Our core innovation is Gaussian-Softmax diffusion, where logits perturbed with Gaussian noise are projected onto the probability simplex via a softmax transformation, facilitating blended class labels for discrete variables. This formulation addresses 2 key challenges, namely, the heterogeneity of primitive parameterizations and the permutation invariance of primitives in CAD sketches. Our approach significantly improves generation quality, reducing Fréchet Inception Distance (FID) from 16.04 to 7.80 and negative log-likelihood (NLL) from 84.8 to 81.33, establishing a new state-of-the-art in CAD sketch generation on the SketchGraphs dataset.

Lay Summary: Generating Computer Aided Design (CAD) blueprints holds great promise for streamlining and democratizing CAD design, similar to images/art with tools like Stable Diffusion. We developed a new AI model for generating CAD sketches in a similar fashion to Stable Diffusion, and introduce a new methodology tailored to categorical data. Our methodology has outperformed all alternatives and brings us one step closer to consumer tools for generating CAD blueprints.

Primary Area: Deep Learning->Generative Models and Autoencoders

Keywords: Diffusion, Discrete Diffusion, CAD, CAD Sketch, Generative, AI, ML, Gaussian-Softmax

Submission Number: 8113