Graph Neural Network– tag –
-
Physics-Informed Graph Neural Network Unlocks Heterogeneous Solid Mechanics: Modeling Deformation and Crack Propagation Without Extensive Data
arXiv Unknown Overview A novel Physics-Informed Graph Neural Network (PI-GNN) has been developed, successfully modeling deformation and crack propagation in heterogeneous solid mechanics with high accuracy. Operating on conforming mesh g... -
arXiv Releases High-Information Dataset ‘MAD-1.6’ for Universal Atomistic Machine Learning, Boosting Predictive Accuracy
arXiv USA Overview The high-quality, high-information dataset 'MAD-1.6' has been released on arXiv to accelerate universal atomistic machine learning. Comprising 362,646 atomic structures across 102 chemical elements, it covers diverse m... -
Google DeepMind GNoME and Microsoft MatterGen Accelerate Materials Discovery, Predicting Millions of Novel Stable Structures
Facebook (MIT DMSE) USA Overview Generative AI models, including Google DeepMind's GNoME and Microsoft's MatterGen, are significantly accelerating materials discovery by proposing novel material combinations and predicting their stabilit... -
Equivariant Graph Neural Network Interatomic Potentials Accelerate Nanomaterials Simulation by 100x
nano-matter.com International Overview Equivariant Graph Neural Network (GNN) interatomic potentials are accelerating nanomaterials research by predicting atomic forces and energies with DFT-comparable accuracy, while reducing computatio... -
GRACE-OFF Achieves High-Precision Machine-Learned Interatomic Potentials for Organic Liquids via GRACE Architecture
Journal of Chemical Theory and Computation | ACS Publications USA Overview This study introduces GRACE-OFF, a machine-learned interatomic potential (MLIP) built on the Graph Atomic Cluster Expansion (GRACE) neural network architecture, d... -
Machine Learning Revolutionizes Magnetic Functional Alloy Design: Generative GNNs and High-Throughput DFT Discover 2D Magnets
AIP Advances USA Overview This article details recent advancements in the design of magnetic functional alloys (MFAs) leveraging machine learning (ML), highlighting how ML has contributed to the design of magnetocaloric, magnetostrictive... -
ACS Publications: PolyCLIP Framework Outperforms Existing Polymer Property Prediction Models by 13.32%, Establishes New Foundation for Multimodal Polymer Informatics
ACS Publications USA Overview ACS Publications announced PolyCLIP, a CLIP-based multimodal framework, achieved up to 13.32% better performance in polymer property prediction than existing unimodal and multimodal models. The study demonst... -
Cypris AI: Generative Models, GNNs, and Autonomous Labs Slash Materials R&D to 1-2 Years; Google DeepMind’s GNoME Predicts 2.4 Million New Materials
Cypris AI USA Overview Cypris AI reports that the integration of generative models, Graph Neural Networks (GNNs), and autonomous labs is drastically cutting materials R&D timelines from 10-20 years to just 1-2 years. Enhanced GNN arc... -
AI Accelerates Material Discovery: Novel GNN Model Generates Electronic Fingerprints for High-Throughput Catalyst & Battery Research
Facebook USA Overview A novel Graph Neural Network (GNN) model has been developed to rapidly generate electronic fingerprints, significantly accelerating the discovery of new materials for catalysts and batteries at a fraction of previou... -
AI Discovers High-Temperature Stable Lead-Free Dielectric Materials Using Multimodal Literature Mining and Physics-Based ML
Facebook USA Overview Researchers leveraged artificial intelligence (AI) to discover novel lead-free dielectric materials with high-temperature stability. The study employed a reverse design approach combining multimodal literature minin...
