The success of Graph Neural Networks (GNNs) in practice has motivated extensive research on their theoretical properties. This includes recent results that characterise node classifiers expressible by GNNs in terms of first order logic. Most of the analysis, however, has been focused on GNNs with fixed number of message-passing iterations (i.e., layers), which cannot realise many simple classifiers such as reachability of a node with a given label. In this paper, we start to fill this gap and study the foundations of GNNs that can perform more than a fixed number of message-passing iterations. We first formalise two generalisations of the basic GNNs: recurrent GNNs (RecGNNs), which repeatedly apply message-passing iterations until the node classifications become stable, and graph-size GNNs (GSGNNs), which exploit a built-in function of the input graph size to decide the number of message-passings. We then formally prove that GNN classifiers are strictly less expressive than RecGNN ones, and RecGNN classifiers are strictly less expressive than GSGNN ones. To get this result, we identify novel semantic characterisations of the three formalisms in terms of suitable variants of bisimulation, which we believe have their own value for our understanding of GNNs. Finally, we prove syntactic logical characterisations of RecGNNs and GSGNNs analogous to the logical characterisation of plain GNNs, where we connect the two formalisms to monadic monotone fixpoint logic---a generalisation of first-order logic that supports recursion.

A seminal theorem by van Benthem characterises the bisimulation-invariant fragment of First Order Logic (FOL) in terms of Modal Logic. Similarly, Janin and Walukiewicz have shown that the bisimulation-invariant fragment of Monadic Second Order Logic (MSO) corresponds precisely to the (modal) mu-calculus. Notably, neither argument immediately carries over when only considering finite structures. In the case of FOL, Rosen provided an alternative proof, which also works over finite structures. However, such characterisations for any logic more expressive than FOL over finite structures have remained long-standing open problems. In this paper, we prove such a characterisation for parameter-free Monadic Least Fixpoint Logic, a well-known logic in between FOL and MSO. In particular, we show that, over finite structures, every two-way bisimulation-invariant formula in this logic is equivalent to a formula in the two-way mu-calculus and vice versa.

Abstract Real-world knowledge graphs (KGs) are usually incomplete—that is, miss some facts representing valid information. So, when applied to such KGs, standard symbolic query engines fail to produce answers that are expected but not logically entailed by the KGs. To overcome this issue, state-of-the-art ML-based approaches first embed KGs and queries into a low-dimensional vector space, and then produce query answers based on the proximity of the candidate entity and the query embeddings in the embedding space. This allows embedding-based approaches to obtain expected answers that are not logically entailed. However, embedding-based approaches are not applicable in the inductive setting, where KG entities (i.e., constants) seen at runtime may differ from those seen during training. In this paper, we propose a novel neuro-symbolic approach to query answering over incomplete KGs applicable in the inductive setting. Our approach first symbolically augments the input KG with facts representing parts of the KG that match query fragments, and then applies a generalisation of the Relational Graph Convolutional Networks (RGCNs) to the augmented KG to produce the predicted query answers. We formally prove that, under reasonable assumptions, our approach can capture an approach based on vanilla RGCNs (and no KG augmentation) using a (often substantially) smaller number of layers. Finally, we empirically validate our theoretical findings by evaluating an implementation of our approach against the RGCN baseline on several dedicated benchmarks.