Category Ranking
98%
Total Visits
921
Avg Visit Duration
2 minutes
Citations
20
Article Abstract
Message passing neural networks have demonstrated significant efficacy in predicting molecular interactions. Introducing equivariant vectorial representations augments expressivity by capturing geometric data symmetries, thereby improving model accuracy. However, two-body bond vectors in opposition may cancel each other out during message passing, leading to the loss of directional information on their shared node. In this study, we develop quivariant -body nteraction works (ENINet) that explicitly integrates = 1 equivariant many-body interactions to enhance directional symmetric information in the message passing scheme. We provided a mathematical analysis demonstrating the necessity of incorporating many-body equivariant interactions and generalized the formulation to -body interactions. Experiments indicate that integrating many-body equivariant representations enhances prediction accuracy across diverse scalar and tensorial quantum chemical properties.
Download full-text PDF |
Source |
|---|---|
| http://www.ncbi.nlm.nih.gov/pmc/articles/PMC12392447 | PMC |
| http://dx.doi.org/10.1021/acs.jctc.5c00466 | DOI Listing |
Publication Analysis
Top Keywords
Similar Publications
Molecule Graph Networks with Many-Body Equivariant Interactions.
J Chem Theory Comput
August 2025
Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa 2778561, Japan.
Message passing neural networks have demonstrated significant efficacy in predicting molecular interactions. Introducing equivariant vectorial representations augments expressivity by capturing geometric data symmetries, thereby improving model accuracy. However, two-body bond vectors in opposition may cancel each other out during message passing, leading to the loss of directional information on their shared node.
View Article and Find Full Text PDFZero shot molecular generation via similarity kernels.
Nat Commun
July 2025
Engineering Laboratory, University of Cambridge, Cambridge, UK.
Generative modelling aims to accelerate the discovery of novel chemicals by directly proposing structures with desirable properties. Recently, score-based, or diffusion, generative models have significantly outperformed previous approaches. Key to their success is the close relationship between the score and physical force, allowing the use of powerful equivariant neural networks.
View Article and Find Full Text PDFMachine Learning Force Fields in Electrochemistry: From Fundamentals to Applications.
ACS Nano
July 2025
Toyota Central R&D Laboratories., Inc., Nagakute 480-1192, Aichi, Japan.
This article reviews the foundations and applications of machine learning force fields (MLFFs) in electrochemistry, highlighting their role as a transformative tool in materials science. We first provide an overview of MLFFs, then discuss their applications in ionics and electrochemical reactions, and finally outline future directions. Most MLFF approaches use invariant or equivariant descriptors derived from body-order expansions to represent many-body atomic interactions.
View Article and Find Full Text PDFSpectroscopy of two-dimensional interacting lattice electrons using symmetry-aware neural backflow transformations.
Commun Phys
January 2025
Institute of Physics, École Polytechnique Fédérale de Lausanne (EPFL), CH-1015 Lausanne, Switzerland.
Neural networks have shown to be a powerful tool to represent the ground state of quantum many-body systems, including fermionic systems. However, efficiently integrating lattice symmetries into neural representations remains a significant challenge. In this work, we introduce a framework for embedding lattice symmetries in fermionic wavefunctions and demonstrate its ability to target both ground states and low-lying excitations.
View Article and Find Full Text PDFX2-GNN: A Physical Message Passing Neural Network with Natural Generalization Ability to Large and Complex Molecules.
J Phys Chem Lett
December 2024
Collaborative Innovation Center of Chemistry for Energy Materials, Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, MOE Key Laboratory of Computational Physical Sciences, Department of Chemistry, Fudan University, Shanghai 200438, China.
Neural network models excel in molecular property predictions but often struggle with generalizing from smaller to larger molecules due to increased structural diversity and complex interactions. To address this, we introduce an E(3) invariant (and equivariant capable) message passing graph neural network (GNN), namely, X2-GNN, that integrates physical insights via atomic orbital overlap integrals and core Hamiltonians. These features provide essential information about bond strength, electron delocalization, and many-body interactions, enhanced by an attention mechanism for improved learning efficiency.
View Article and Find Full Text PDF