Distributionally Robust Semi-Supervised Learning Over Graphs
This addresses robustness issues in graph neural networks for network science applications, but it is incremental as it builds on existing distributionally robust optimization methods.
The paper tackles the problem of semi-supervised learning over graphs when nodal attributes are uncertain due to mismatches between training and testing distributions or noisy measurements, by developing a distributionally robust learning framework that minimizes worst-case expected loss within a Wasserstein ball, resulting in a tractable and efficiently solvable method.
Semi-supervised learning (SSL) over graph-structured data emerges in many network science applications. To efficiently manage learning over graphs, variants of graph neural networks (GNNs) have been developed recently. By succinctly encoding local graph structures and features of nodes, state-of-the-art GNNs can scale linearly with the size of graph. Despite their success in practice, most of existing methods are unable to handle graphs with uncertain nodal attributes. Specifically whenever mismatches between training and testing data distribution exists, these models fail in practice. Challenges also arise due to distributional uncertainties associated with data acquired by noisy measurements. In this context, a distributionally robust learning framework is developed, where the objective is to train models that exhibit quantifiable robustness against perturbations. The data distribution is considered unknown, but lies within a Wasserstein ball centered around empirical data distribution. A robust model is obtained by minimizing the worst expected loss over this ball. However, solving the emerging functional optimization problem is challenging, if not impossible. Advocating a strong duality condition, we develop a principled method that renders the problem tractable and efficiently solvable. Experiments assess the performance of the proposed method.