# Masked BRep Autoencoder via Hierarchical Graph Transformer

arXiv: https://arxiv.org/abs/2603.14927 · HTML: https://mlsanity.com/paper/2603.14927/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 39 references that are in the map.
- First posted on arXiv 16 Mar 2026.

## Builds on (most cited first, 18 of 18)

- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Rethinking Graph Transformers with Spectral Attention](https://mlsanity.com/paper/2106.03893/) (2021)
- [Do Transformers Really Perform Bad for Graph Representation?](https://mlsanity.com/paper/2106.05234/) (2021)
- [Representing Long-Range Context for Graph Neural Networks with Global Attention](https://mlsanity.com/paper/2201.08821/) (2022)
- [Graph-Bert: Only Attention is Needed for Learning Graph Representations](https://mlsanity.com/paper/2001.05140/) (2020)
- [GraphiT: Encoding Graph Structure in Transformers](https://mlsanity.com/paper/2106.05667/) (2021)
- [SGFormer: Simplifying and Empowering Transformers for Large-Graph Representations](https://mlsanity.com/paper/2306.10759/) (2023)
- [BrepGen: A B-rep Generative Diffusion Model with Structured Latent Geometry](https://mlsanity.com/paper/2401.15563/) (2024)
- [BRepNet: A topological message passing system for solid models](https://mlsanity.com/paper/2104.00706/) (2021)
- [JoinABLe: Learning Bottom-up Assembly of Parametric CAD Joints](https://mlsanity.com/paper/2111.12772/) (2021)
- [Self-Supervised Representation Learning for CAD](https://mlsanity.com/paper/2210.10807/) (2022)
- [Evaluating Deep Graph Neural Networks](https://mlsanity.com/paper/2108.00955/) (2021)
- [Bringing Attention to CAD: Boundary Representation Learning via Transformer](https://mlsanity.com/paper/2504.07134/) (2025)
- [BRepFormer: Transformer-Based B-rep Geometric Feature Recognition](https://mlsanity.com/paper/2504.07378/) (2025)
- [Representing 3D Shapes With 64 Latent Vectors for 3D Diffusion Models](https://mlsanity.com/paper/2503.08737/) (2025)
- [Implicit Conversion of Manifold B-Rep Solids by Neural Halfspace Representation](https://mlsanity.com/paper/2209.10191/) (2022)
- [BRepMAE: Self-Supervised Masked BRep Autoencoders for Machining Feature Recognition](https://mlsanity.com/paper/2602.22701/) (2026)
- [Material Prediction for Design Automation Using Graph Representation Learning](https://mlsanity.com/paper/2209.12793/) (2022)

## Built on by (most cited first, 3 of 3)

- [Masked Topology Modeling for Self-Supervised Learning on Parametric CAD](https://mlsanity.com/paper/2607.20642/) (2026)
- [Learn the Solid, Not the File: Canonical Inputs for Neural Networks on CAD Boundary Representations](https://mlsanity.com/paper/2609.11573/) (2026)
- [CNCGEN: A Dataset and Framework for Machining Process Planning and Toolpath Generation from B-rep Models](https://mlsanity.com/paper/2609.39738/) (2026)

## Abstract (arXiv)

We introduce a novel self-supervised learning framework that automatically learns representations from input computer-aided design (CAD) models for downstream tasks, including part classification, modeling segmentation, and machining feature recognition. To train our network, we construct a large-scale, unlabeled dataset of boundary representation (BRep) models. The success of our algorithm relies on two keycomponents. The first is a masked graph autoencoder that reconstructs randomly masked geometries and attributes of BReps for representation learning to enhance the generalization. The second is a hierarchical graph Transformer architecture that elegantly fuses global and local learning by a cross-scale mutual attention block to model long-range geometric dependencies and a graph neural network block to aggregate local topological information. After training the autoencoder, we replace its decoder with a task-specific network trained on a small amount of labeled data for downstream tasks. We conduct experiments on various tasks and achieve high performance, even with a small amount of labeled data, demonstrating the practicality and generalizability of our model. Compared to other methods, our model performs significantly better on downstream tasks with the same amount of training data, particularly when the training data is very limited.