Timezone: »

 
Oral
Equivariant Diffusion for Molecule Generation in 3D
Emiel Hoogeboom · Victor Garcia Satorras · Clément Vignac · Max Welling

Tue Jul 19 08:05 AM -- 08:25 AM (PDT) @ Room 310
in Deep Learning: Generative Models/Autoencoders »

This work introduces a diffusion model for molecule generation in 3D that is equivariant to Euclidean transformations. Our E(3) Equivariant Diffusion Model (EDM) learns to denoise a diffusion process with an equivariant network that jointly operates on both continuous (atom coordinates) and categorical features (atom types). In addition, we provide a probabilistic analysis which admits likelihood computation of molecules using our model. Experimentally, the proposed method significantly outperforms previous 3D molecular generative methods regarding the quality of generated samples and the efficiency at training time.

Keywords: APP: Chemistry and Drug Discovery · DL: Generative Models and Autoencoders

Author Information

Emiel Hoogeboom (University of Amsterdam)
Victor Garcia Satorras (Microsoft)
Clément Vignac (EPFL)
Max Welling (University of Amsterdam & Qualcomm)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors