-
Exploring Chemical Space with Score-based Out-of-distribution Generation
Score-based generative models have shown promise in molecule generation, but often struggle to create truly novel candidates beyond the training distribution. MOOD (Molecular Out-Of-distribution Diffusion) is a score-based diffusion framework that enables controllable exploration of out-of-distribution chemical space without incurring additional computational costs. By integrating a property prediction network into the reverse-time SDE, MOOD effectively guides the generation toward molecules with desired novel traits such as high binding affinity, drug-likeness, and synthesizability.
-
Score-based Generative Modeling of Graphs via the System of Stochastic Differential Equations
Score-based generative models have achieved strong results in image and molecular generation, but applying them to graph-structured data remains challenging due to the complex interplay between node features and graph topology. In this blog, we explore GDSS (Graph Diffusion via the System of Stochastic differential equations), a model that tackles this problem by jointly modeling the evolution of node attributes and adjacency matrices through a system of coupled stochastic differential equations. Using a permutation-equivariant graph neural network and score matching, GDSS generates graphs that maintain both structural validity and semantic coherence.