Semi-Equivariant Conditional Normalizing Flows

Abstract

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.

Cite

Text

Rozenberg and Freedman. "Semi-Equivariant Conditional Normalizing Flows." ICLR 2023 Workshops: Physics4ML, 2023.

Markdown

[Rozenberg and Freedman. "Semi-Equivariant Conditional Normalizing Flows." ICLR 2023 Workshops: Physics4ML, 2023.](https://mlanthology.org/iclrw/2023/rozenberg2023iclrw-semiequivariant/)

BibTeX

@inproceedings{rozenberg2023iclrw-semiequivariant,
  title     = {{Semi-Equivariant Conditional Normalizing Flows}},
  author    = {Rozenberg, Eyal and Freedman, Daniel},
  booktitle = {ICLR 2023 Workshops: Physics4ML},
  year      = {2023},
  url       = {https://mlanthology.org/iclrw/2023/rozenberg2023iclrw-semiequivariant/}
}