Skip to yearly menu bar Skip to main content


Learning Discrete Structures for Graph Neural Networks

Luca Franceschi · Mathias Niepert · Massimiliano Pontil · Xiao He

Pacific Ballroom #177

Keywords: [ Semi-supervised learning ] [ Optimization ] [ Non-convex Optimization ] [ Networks and Relational Learning ] [ Algorithms ]


Graph neural networks (GNNs) are a popular class of machine learning models that have been successfully applied to a range of problems. Their major advantage lies in their ability to explicitly incorporate a sparse and discrete dependency structure between data points. Unfortunately, GNNs can only be used when such a graph-structure is available. In practice, however, real-world graphs are often noisy and incomplete or might not be available at all. With this work, we propose to jointly learn the graph structure and the parameters of graph convolutional networks (GCNs) by approximately solving a bilevel program that learns a discrete probability distribution on the edges of the graph.
This allows one to apply GCNs not only in scenarios where the given graph is incomplete or corrupted but also in those where a graph is not available. We conduct a series of experiments that analyze the behavior of the proposed method and demonstrate that it outperforms related methods by a significant margin.

Live content is unavailable. Log in and register to view live content