Siamese Foundation Models for Crystal Structure Prediction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wu, Liming, Huang, Wenbing, Jiao, Rui, Huang, Jianxing, Liu, Liwei, Zhou, Yipeng, Sun, Hao, Liu, Yang, Sun, Fuchun, Ren, Yuxiang, Wen, Jirong
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917404936765440
author Wu, Liming
Huang, Wenbing
Jiao, Rui
Huang, Jianxing
Liu, Liwei
Zhou, Yipeng
Sun, Hao
Liu, Yang
Sun, Fuchun
Ren, Yuxiang
Wen, Jirong
author_facet Wu, Liming
Huang, Wenbing
Jiao, Rui
Huang, Jianxing
Liu, Liwei
Zhou, Yipeng
Sun, Hao
Liu, Yang
Sun, Fuchun
Ren, Yuxiang
Wen, Jirong
contents Predicting crystal structures from chemical compositions is a fundamental challenge in materials discovery, complicated by complex 3D geometries that distinguish it from fields like protein folding. Here, we present Diffusion-based Crystal Omni (DAO), a pretrain-finetune framework for crystal structure prediction integrating two Siamese foundation models: a structure generator and an energy predictor. The generator is pretrained via a two-stage pipeline on a vast dataset of stable and unstable structures, leveraging the predictor to relax unstable configurations and guide the generative sampling. Across two well-known benchmarks, pretraining significantly enhances performance across multiple backbone architectures. Ablation studies confirm that the synergy between the generator and predictor mutually benefits both components. We further validate DAO on three real-world superconductors ($\text{Cr}_6\text{Os}_2$, $\text{Zr}_{16}\text{Rh}_8\text{O}_4$, and $\text{Zr}_{16}\text{Pd}_8\text{O}_4$) typically inaccessible to conventional computation. For $\text{Cr}_6\text{Os}_2$, DAO achieves a 100\% match rate with experimental references and an atomic-position error of 0.0012 under 20-shot generation, performing over 2000$\times$ faster per iteration than DFT-based structure predictors. These compelling results collectively highlight the potential of our approach for advancing materials science research.
format Preprint
id arxiv_https___arxiv_org_abs_2503_10471
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Siamese Foundation Models for Crystal Structure Prediction
Wu, Liming
Huang, Wenbing
Jiao, Rui
Huang, Jianxing
Liu, Liwei
Zhou, Yipeng
Sun, Hao
Liu, Yang
Sun, Fuchun
Ren, Yuxiang
Wen, Jirong
Materials Science
Artificial Intelligence
Predicting crystal structures from chemical compositions is a fundamental challenge in materials discovery, complicated by complex 3D geometries that distinguish it from fields like protein folding. Here, we present Diffusion-based Crystal Omni (DAO), a pretrain-finetune framework for crystal structure prediction integrating two Siamese foundation models: a structure generator and an energy predictor. The generator is pretrained via a two-stage pipeline on a vast dataset of stable and unstable structures, leveraging the predictor to relax unstable configurations and guide the generative sampling. Across two well-known benchmarks, pretraining significantly enhances performance across multiple backbone architectures. Ablation studies confirm that the synergy between the generator and predictor mutually benefits both components. We further validate DAO on three real-world superconductors ($\text{Cr}_6\text{Os}_2$, $\text{Zr}_{16}\text{Rh}_8\text{O}_4$, and $\text{Zr}_{16}\text{Pd}_8\text{O}_4$) typically inaccessible to conventional computation. For $\text{Cr}_6\text{Os}_2$, DAO achieves a 100\% match rate with experimental references and an atomic-position error of 0.0012 under 20-shot generation, performing over 2000$\times$ faster per iteration than DFT-based structure predictors. These compelling results collectively highlight the potential of our approach for advancing materials science research.
title Siamese Foundation Models for Crystal Structure Prediction
topic Materials Science
Artificial Intelligence
url https://arxiv.org/abs/2503.10471