# Data-driven discovery of novel 2D materials by deep generative models

arXiv: https://arxiv.org/abs/2206.12159 · HTML: https://mlsanity.com/paper/2206.12159/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- Builds on 6 of its 52 references that are in the map.
- First posted on arXiv 24 Jun 2022.

## Builds on (most cited first, 6 of 6)

- [Generative Modeling by Estimating Gradients of the Data Distribution](https://mlsanity.com/paper/1907.05600/) (2019)
- [Auto-Encoding Variational Bayes](https://mlsanity.com/paper/1312.6114/) (2013)
- [Crystal Diffusion Variational Autoencoder for Periodic Material Generation](https://mlsanity.com/paper/2110.06197/) (2021)
- [An invertible crystallographic representation for general inverse design of inorganic crystals with targeted properties](https://mlsanity.com/paper/2005.07609/) (2020)
- [AiiDA: Automated Interactive Infrastructure and Database for Computational Science](https://mlsanity.com/paper/1504.01163/) (2015)
- [Computational discovery of new 2D materials using deep learning generative models](https://mlsanity.com/paper/2012.09314/) (2020)

## Built on by (most cited first, 18 of 18)

- [MOFDiff: Coarse-grained Diffusion for Metal-Organic Framework Design](https://mlsanity.com/paper/2310.10732/) (2023)
- [Kinetic Langevin Diffusion for Crystalline Materials Generation](https://mlsanity.com/paper/2507.03602/) (2025)
- [Discovery of 2D materials using Transformer Network based Generative Design](https://mlsanity.com/paper/2301.05824/) (2023)
- [Artificial Intelligence in Materials Science and Engineering: Current Landscape, Key Challenges, and Future Trajectorie](https://mlsanity.com/paper/2601.12554/) (2026)
- [Leveraging Language Representation for Material Recommendation, Ranking, and Exploration](https://mlsanity.com/paper/2305.01101/) (2023)
- [Generative Design of inorganic compounds using deep diffusion language models](https://mlsanity.com/paper/2310.00475/) (2023)
- [Bayesian optimization of atomic structures with prior probabilities from universal interatomic potentials](https://mlsanity.com/paper/2408.15590/) (2024)
- [Discovery of 2D Materials via Symmetry-Constrained Diffusion Model](https://mlsanity.com/paper/2412.18414/) (2024)
- [SLayerGen: a Crystal Generative Model for all Space and Layer Groups](https://mlsanity.com/paper/2605.08262/) (2026)
- [Contrastive Learning of English Language and Crystal Graphs for Multimodal Representation of Materials Knowledge](https://mlsanity.com/paper/2502.16451/) (2025)
- [Gradient-based grand canonical optimization enabled by graph neural networks with fractional atomic existence](https://mlsanity.com/paper/2507.19438/) (2025)
- [Rapid Discovery of Graphene Nanocrystals Using DFT and Bayesian Optimization with Neural Network Kernel](https://mlsanity.com/paper/2208.07612/) (2022)
- [Combining unsupervised and supervised learning in microscopy enables defect analysis of a full 4H-SiC wafer](https://mlsanity.com/paper/2402.13353/) (2024)
- [Unified machine-learning framework for property prediction and time-evolution simulation of strained alloy microstructure](https://mlsanity.com/paper/2507.21760/) (2025)
- [Materium: An Autoregressive Approach for Material Generation](https://mlsanity.com/paper/2512.07486/) (2025)
- [A Critical Examination of Active Learning Workflows in Materials Science](https://mlsanity.com/paper/2601.05946/) (2026)
- [Neural surrogates for crystal growth dynamics with variable supersaturation: explicit vs. implicit conditioning](https://mlsanity.com/paper/2604.21753/) (2026)
- [InvDesMobility: a reliability-gated first-principles feedback framework for closed-loop materials discovery](https://mlsanity.com/paper/2606.16133/) (2026)

## Abstract (arXiv)

Efficient algorithms to generate candidate crystal structures with good stability properties can play a key role in data-driven materials discovery. Here we show that a crystal diffusion variational autoencoder (CDVAE) is capable of generating two-dimensional (2D) materials of high chemical and structural diversity and formation energies mirroring the training structures. Specifically, we train the CDVAE on 2615 2D materials with energy above the convex hull Delta Hhull< 0.3 eV/atom, and generate 5003 materials that we relax using density functional theory (DFT). We also generate 14192 new crystals by systematic element substitution of the training structures. We find that the generative model and lattice decoration approach are complementary and yield materials with similar stability properties but very different crystal structures and chemical compositions. In total we find 11630 predicted new 2D materials, where 8599 of these have Delta Hhull< 0.3 eV/atom as the seed structures, while 2004 are within 50 meV of the convex hull and could potentially be synthesized. The relaxed atomic structures of all the materials are available in the open Computational 2D Materials Database (C2DB). Our work establishes the CDVAE as an efficient and reliable crystal generation machine, and significantly expands the space of 2D materials.