Equivariant Graph Neural Networks (EGNNs) are a type of neural network designed to preserve symmetries in data, particularly for graph-structured input. These networks are “equivariant,” meaning they maintain the relationship between input and output when transformations (such as rotations or reflections) are applied. EGNNs are especially useful in tasks where the data’s geometric properties need to be preserved, such as in 3D molecular modeling or physical simulations. They enable more accurate predictions while respecting the underlying symmetries of the data.

Posts

Learning Disentangled Equivariant Representation for Explicitly Controllable 3D Molecule Generation

/in /by

We consider the conditional generation of 3D drug-like molecules with explicit control over molecular properties such as drug-like properties (e.g., Quantitative Estimate of Druglikenessor Synthetic Accessibility score) and effectively binding to specific protein sites. To tackle this problem, we propose an E(3)-equivariant Wasserstein autoencoder and factorize thelatent space of our generative model into two disentangled aspects: molecular properties and the remaining structural context of 3D molecules. Our model ensures explicit control over these molecular attributes while maintaining equivariance of coordinate representation and invariance of data likelihood. Furthermore, we introduce a novel alignment-based coordinate loss to adapt equivariant networks for auto-regressive denovo 3D molecule generation from scratch. Extensive experiments validate our model’s effectiveness on property-guidedand context-guided molecule generation, both for de-novo 3D molecule design and structure-based drug discovery against protein targets.

Read more