Condensed Matter > Soft Condensed Matter
[Submitted on 8 Jul 2024 (v1), last revised 9 Jul 2024 (this version, v2)]
Title:Enhancing the Prediction of Glass Dynamics by Incorporating the Direction of Deviation from Equilibrium Positions
View PDF HTML (experimental)Abstract:Elucidating the intricate relationship between the structure and dynamics in the context of the glass transition has been a persistent challenge. Machine learning (ML) has emerged as a pivotal tool, offering novel pathways to predict dynamic behaviors from structural descriptors. Notably, recent research has highlighted that the distance between the initial particle positions between the equilibrium positions substantially enhances the prediction of glassy dynamics. However, these methodologies have been limited in their ability to capture the directional aspects of these deviations from the equilibrium positions, which are crucial for a comprehensive understanding of the complex particle interactions within the cage dynamics. Therefore, this paper introduces a novel structural parameter: the vectorial displacement of particles from their initial configuration to their equilibrium positions. Recognizing the inadequacy of current ML models in effectively handling such vectorial parameters, we have developed an Equivariance-Constrained Invariant Graph Neural Network (EIGNN). This innovative model not only bolsters the descriptive capacity of conventional rotation-invariant models but also streamlines the computational demands associated with rotation-equivariant graph neural networks. Our rigorous experimental validation on 3D glassy system from GlassBench dataset has yielded compelling evidence that the EIGNN model significantly enhance the correlation between structural representation and dynamic properties.
Submission history
From: Xiao Jiang [view email][v1] Mon, 8 Jul 2024 16:48:26 UTC (1,297 KB)
[v2] Tue, 9 Jul 2024 18:41:44 UTC (1,814 KB)
Current browse context:
cond-mat.soft
Change to browse by:
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
IArxiv Recommender
(What is IArxiv?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.