Autoregressive Generative Neural Networks for the Inverse Design of 3d Molecular Structures
Autoregressive Generative Neural Networks for the Inverse Design of 3d Molecular Structures The rational design of molecules with desired properties is a long-standing challenge in chemistry. Deep learning has proven to yield fast and accurate predictions of quantum-chemical properties to accelerate the discovery of novel molecules and materials. As an exhaustive exploration of the vast chemical space is still infeasible, we require generative models that guide our search towards systems with desired properties. While graph-based models have previously been proposed, they are restricted by a lack of spatial information such that they are unable to recognize spatial isomerism and non-bonded interactions. In this thesis, we develop a conditional generative neural network for 3d molecules that respects the symmetries inherent to such structures. It enables the sampling of novel molecular structures from conditional distributions, even in domains where reference calculations are sparse. We demonstrate the utility of our method for inverse design by generating molecules with specified motifs or composition, discovering particularly stable molecules, and jointly targeting multiple quantum chemical properties beyond the training regime. Finally, we present an optimized reference implementation in an open source software package and provide comprehensive instructions on its usage to make the method widely accessible.
