top of page
Semi-Equivariant Conditional Normalizing Flows
E. Rozenberg and D. Freedman
International Conference on Learning Representations (ICLR) Workshop on Physics for Machine Learning Workshop (Physics4ML), 2023
We study the problem of learning conditional distributions of the form $p(G|\hat{G})$, where $G$ and $\hat{G}$ are two 3D graphs, using continuous normalizing flows. We derive a semi-equivariance condition on the flow which ensures that conditional invariance to rigid motions holds. We demonstrate the effectiveness of the technique in the molecular setting of receptor-aware ligand generation.
bottom of page