Webpages tagged with «Graph Neural Networks»

Two interesting connections have been recently discovered between the symbolic and sub-symbolic AI paradigms for data management, specifically between computational logics and Graph Neural Networks (GNNs). First, it was shown that certain GNN architectures are able to express some fragments of first-order logic, but fail to do so with others due to lack of enough expressive power. Here, expressivity of a logic fragment by a GNN architecture is theoretical: it means that every function on graphs given by a formula in the logical fragment can be written as a GNN model in the architecture. Yet, the fact that there exists such a GNN model does not immediately mean that we can train that model in practice with a satisfactory precision. Thus, second, it has been also shown empirically, that for some GNN architectures can be trained to learn certain theoretically expressible logical functions, but cannot do this for others (at least, using standard training techniques).

There are many GNN architectures, and these two connections are not yet known for all of them. In this project, the student will investigate one or both for architectures that are not yet covered.