Structural Constraint Integration in Generative Model for Discovery of Quantum Material Candidates
Authors:
Ryotaro Okabe,
Mouyang Cheng,
Abhijatmedhi Chotrattanapituk,
Nguyen Tuan Hung,
Xiang Fu,
Bowen Han,
Yao Wang,
Weiwei Xie,
Robert J. Cava,
Tommi S. Jaakkola,
Yongqiang Cheng,
Mingda Li
Abstract:
Billions of organic molecules are known, but only a tiny fraction of the functional inorganic materials have been discovered, a particularly relevant problem to the community searching for new quantum materials. Recent advancements in machine-learning-based generative models, particularly diffusion models, show great promise for generating new, stable materials. However, integrating geometric patt…
▽ More
Billions of organic molecules are known, but only a tiny fraction of the functional inorganic materials have been discovered, a particularly relevant problem to the community searching for new quantum materials. Recent advancements in machine-learning-based generative models, particularly diffusion models, show great promise for generating new, stable materials. However, integrating geometric patterns into materials generation remains a challenge. Here, we introduce Structural Constraint Integration in the GENerative model (SCIGEN). Our approach can modify any trained generative diffusion model by strategic masking of the denoised structure with a diffused constrained structure prior to each diffusion step to steer the generation toward constrained outputs. Furthermore, we mathematically prove that SCIGEN effectively performs conditional sampling from the original distribution, which is crucial for generating stable constrained materials. We generate eight million compounds using Archimedean lattices as prototype constraints, with over 10% surviving a multi-staged stability pre-screening. High-throughput density functional theory (DFT) on 26,000 survived compounds shows that over 50% passed structural optimization at the DFT level. Since the properties of quantum materials are closely related to geometric patterns, our results indicate that SCIGEN provides a general framework for generating quantum materials candidates.
△ Less
Submitted 5 July, 2024;
originally announced July 2024.
Virtual Node Graph Neural Network for Full Phonon Prediction
Authors:
Ryotaro Okabe,
Abhijatmedhi Chotrattanapituk,
Artittaya Boonkird,
Nina Andrejevic,
Xiang Fu,
Tommi S. Jaakkola,
Qichen Song,
Thanh Nguyen,
Nathan Drucker,
Sai Mu,
Bolin Liao,
Yongqiang Cheng,
Mingda Li
Abstract:
The structure-property relationship plays a central role in materials science. Understanding the structure-property relationship in solid-state materials is crucial for structure design with optimized properties. The past few years witnessed remarkable progress in correlating structures with properties in crystalline materials, such as machine learning methods and particularly graph neural network…
▽ More
The structure-property relationship plays a central role in materials science. Understanding the structure-property relationship in solid-state materials is crucial for structure design with optimized properties. The past few years witnessed remarkable progress in correlating structures with properties in crystalline materials, such as machine learning methods and particularly graph neural networks as a natural representation of crystal structures. However, significant challenges remain, including predicting properties with complex unit cells input and material-dependent, variable-length output. Here we present the virtual node graph neural network to address the challenges. By developing three types of virtual node approaches - the vector, matrix, and momentum-dependent matrix virtual nodes, we achieve direct prediction of $Γ$-phonon spectra and full dispersion only using atomic coordinates as input. We validate the phonon bandstructures on various alloy systems, and further build a $Γ$-phonon database containing over 146,000 materials in the Materials Project. Our work provides an avenue for rapid and high-quality prediction of phonon spectra and bandstructures in complex materials, and enables materials design with superior phonon properties for energy applications. The virtual node augmentation of graph neural networks also sheds light on designing other functional properties with a new level of flexibility.
△ Less
Submitted 5 January, 2023;
originally announced January 2023.
