Diff-Dock-Chem
[PYTORCH]A score-based diffusion model for molecular docking. Achieved state-of-the-art results in blind docking benchmarks by treating ligand positioning as a generative process on a Riemannian manifold.
View Repository ->
> DEPARTMENT OF CHEMICAL ENGINEERING
> MASSACHUSETTS INSTITUTE OF TECHNOLOGY
> MASSACHUSETTS INSTITUTE OF TECHNOLOGY
GENERATIVE
MATTER
Designing molecular structures and exploring chemical space through deep generative models and geometric deep learning.
STATUS: RESEARCHING
LOC: CAMBRIDGE, MA
SELECTED DEPLOYMENTS
[01-INDEX]Latent-Mol
[VAE]Exploring latent space optimization for small molecule drug discovery using variational autoencoders.
Coming Soon
Protein-GNN
[GRAPH-ML]Graph Neural Networks applied to protein-protein interaction prediction. Utilizing geometric priors to improve generalization on unseen protein folds.
Read Paper ->
Rxn-Transformer
[NLP]Adapting Large Language Models for chemical reaction prediction. Treating SMILES strings as natural language tokens to predict synthesis pathways.
Interactive Demo ->