Awesome AI4CAD Hub Papers · Datasets · Projects
← Back to papers

Mamba-CAD: State Space Model For 3D Computer-Aided Design Generative Modeling

arXiv 2026 67.3 method

TLDR

Mamba-CAD uses a state space model for self-supervised generative modeling of long parametric CAD sequences, with a new dataset of 77,078 models.

Reasoning

Strengths include addressing long sequence modeling in CAD and introducing a new dataset; weaknesses are limited detail on evaluation metrics and lack of comparison to existing methods in the abstract.

Read-first score

Read-first score 67.3, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 68.

Recency 8%
100

Uses a gentle age decay so recent papers surface without erasing older foundations. 2026

Reproducibility 25%
85

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=True; dataset=False; markers=code,dataset,github

Methodology quality 25%
80

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=dataset,evaluation,experiment,metric

Topical relevance 42%
42.5

Uses existing LLM keyword relevance scores normalized to 0-100. AI for CAD,computer-aided design,neural CAD,generative CAD,parametric CAD,B-Rep,boundary representation,constructive solid geometry,CSG,sketch extrusion,CAD generation,CAD reconstruction,text-to-CAD,image-to-CAD,point cloud to CAD,CAD program

Field roles

FrontierMethodology anchorReproducibility anchor

Rank sensitivity

Stability: volatile; rank range: 23.

Keyword Scores

computer-aided design
10
generative CAD
10
parametric CAD
10
CAD generation
10
neural CAD
8
CAD reconstruction
8
AI for CAD
7
CAD program
5
B-Rep
0
boundary representation
0
constructive solid geometry
0
CSG
0
sketch extrusion
0
text-to-CAD
0
image-to-CAD
0
point cloud to CAD
0

Deep Analysis

Innovations

  • Mamba-CAD: a self-supervised generative model for complex CAD models using Mamba architecture to handle longer parametric CAD sequences.
  • Encoder-decoder framework with Mamba, pre-trained via CAD reconstruction to learn latent representations, then using a GAN to generate fake representations decoded to CAD sequences.
  • Creation of a new dataset of 77,078 CAD models with longer parametric CAD sequences.

Methodology

Mamba-CAD uses a Mamba-based encoder-decoder pre-trained on a CAD reconstruction task to learn latent representations. A generative adversarial network then generates fake latent representations, which are decoded back into parametric CAD sequences. A new dataset of 77,078 CAD models with longer sequences is used for training.

Key Results

The model shows effectiveness under various evaluation metrics, particularly in generating longer valid parametric CAD sequences.

Tags

CVAI