About this document
Scalable GNNs for Materials Prediction by wyichen751 is a document available to read on EtoBox.
The article presents DeeperGATGNN, a scalable graph neural network model designed for high-performance materials property prediction, which utilizes a global attention mechanism, differentiable group normalization, and residual connections to overcome the over-smoothing issue prevalent in existing models. This model achieves state-of-the-art performance on five out of six benchmark datasets and can scale to over 30 layers without significant performance loss, making it suitable for large-scale applications.
- Author
- wyichen751
- Language
- EN