Graph Convolutional Policy Network for Goal-Directed Molecular Graph Generation
This addresses the challenge of discovering novel molecules with specific properties for applications in chemistry and drug discovery, representing a strong domain-specific advance.
The paper tackled the problem of generating molecular graphs that optimize desired properties while obeying complex, non-differentiable rules like chemical valency, achieving a 61% improvement in chemical property optimization and a 184% improvement in constrained property optimization over state-of-the-art baselines.
Generating novel graph structures that optimize given objectives while obeying some given underlying rules is fundamental for chemistry, biology and social science research. This is especially important in the task of molecular graph generation, whose goal is to discover novel molecules with desired properties such as drug-likeness and synthetic accessibility, while obeying physical laws such as chemical valency. However, designing models to find molecules that optimize desired properties while incorporating highly complex and non-differentiable rules remains to be a challenging task. Here we propose Graph Convolutional Policy Network (GCPN), a general graph convolutional network based model for goal-directed graph generation through reinforcement learning. The model is trained to optimize domain-specific rewards and adversarial loss through policy gradient, and acts in an environment that incorporates domain-specific rules. Experimental results show that GCPN can achieve 61% improvement on chemical property optimization over state-of-the-art baselines while resembling known molecules, and achieve 184% improvement on the constrained property optimization task.