Latent Diffusion Pretraining for Crystal Property Prediction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Mukherjee, Shrimon, Das, Kishalay, Basuchowdhuri, Partha, Goyal, Pawan, Ganguly, Niloy
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914620528132096
author Mukherjee, Shrimon
Das, Kishalay
Basuchowdhuri, Partha
Goyal, Pawan
Ganguly, Niloy
author_facet Mukherjee, Shrimon
Das, Kishalay
Basuchowdhuri, Partha
Goyal, Pawan
Ganguly, Niloy
contents Fast and accurate prediction of crystal properties is a central challenge in new materials design. Graph neural networks and Transformer-based models have emerged as powerful tools for this task due to their ability to encode the local structural environment of atoms within a crystal. However, these models are data-hungry, and in practice, labeled data for crystal properties are scarce. Pretraining-finetuning strategies, particularly those based on diffusion models, have shown promise in addressing these limitations. In this work, we introduce a novel latent diffusion based pretraining framework, CrysLDNet, designed to mitigate data scarcity. Our approach integrates a Variational Autoencoder (VAE) with a diffusion model during the pretraining stage. The VAE encoder maps 3D crystal structures into a smooth latent space within which the diffusion process is applied. This latent diffusion pretraining enables the graph encoder to effectively capture structural and chemical semantics from large-scale unlabeled data, which can then be finetuned for specific property prediction tasks. Comprehensive experiments on popular DFT datasets for property prediction reveal that CrysLDNet significantly outperforms both training-from-scratch and pretrained baselines, with improvements of 4.26% and 4.90% on the JARVIS and MP datasets, respectively. Additionally, the learned representations remain robust in sparse-data conditions and are expressive enough to correct DFT errors when finetuned with limited experimental data. Code is available at: https://github.com/shrimonmuke0202/CrysLDNet.git.
format Preprint
id arxiv_https___arxiv_org_abs_2606_00776
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Latent Diffusion Pretraining for Crystal Property Prediction
Mukherjee, Shrimon
Das, Kishalay
Basuchowdhuri, Partha
Goyal, Pawan
Ganguly, Niloy
Machine Learning
Fast and accurate prediction of crystal properties is a central challenge in new materials design. Graph neural networks and Transformer-based models have emerged as powerful tools for this task due to their ability to encode the local structural environment of atoms within a crystal. However, these models are data-hungry, and in practice, labeled data for crystal properties are scarce. Pretraining-finetuning strategies, particularly those based on diffusion models, have shown promise in addressing these limitations. In this work, we introduce a novel latent diffusion based pretraining framework, CrysLDNet, designed to mitigate data scarcity. Our approach integrates a Variational Autoencoder (VAE) with a diffusion model during the pretraining stage. The VAE encoder maps 3D crystal structures into a smooth latent space within which the diffusion process is applied. This latent diffusion pretraining enables the graph encoder to effectively capture structural and chemical semantics from large-scale unlabeled data, which can then be finetuned for specific property prediction tasks. Comprehensive experiments on popular DFT datasets for property prediction reveal that CrysLDNet significantly outperforms both training-from-scratch and pretrained baselines, with improvements of 4.26% and 4.90% on the JARVIS and MP datasets, respectively. Additionally, the learned representations remain robust in sparse-data conditions and are expressive enough to correct DFT errors when finetuned with limited experimental data. Code is available at: https://github.com/shrimonmuke0202/CrysLDNet.git.
title Latent Diffusion Pretraining for Crystal Property Prediction
topic Machine Learning
url https://arxiv.org/abs/2606.00776