Graph Random Neural Features for Distance-Preserving Graph Representations
This work addresses the need for efficient and accurate graph representations in machine learning, though it appears incremental as it builds on existing graph neural network and embedding techniques.
The authors tackled the problem of embedding graph-structured data into real vectors while preserving metric distances, introducing Graph Random Neural Features (GRNF) as a method that ensures graph isomorphism handling and approximates distances with theoretical guarantees.
We present Graph Random Neural Features (GRNF), a novel embedding method from graph-structured data to real vectors based on a family of graph neural networks. The embedding naturally deals with graph isomorphism and preserves the metric structure of the graph domain, in probability. In addition to being an explicit embedding method, it also allows us to efficiently and effectively approximate graph metric distances (as well as complete kernel functions); a criterion to select the embedding dimension trading off the approximation accuracy with the computational cost is also provided. GRNF can be used within traditional processing methods or as a training-free input layer of a graph neural network. The theoretical guarantees that accompany GRNF ensure that the considered graph distance is metric, hence allowing to distinguish any pair of non-isomorphic graphs.