Method for providing an agent for creating a graph neural network architecture and method for creating, by an agent, a graph neural network architecture

This invention describes an AI program, called an agent, that learns to design other AI programs called Graph Neural Networks (GNNs). These GNNs are specifically built to predict things within complex systems, like how a city's traffic flows or a biological network behaves. The agent learns by studying various complex systems described in a standardized way, potentially as knowledge graphs, and then uses this knowledge to create a suitable GNN for a new, unknown complex system.