Private Graphon Estimation for Sparse Graphs
This work addresses privacy concerns in network analysis for applications like social networks, offering a solution that protects individual identities while enabling accurate graph structure estimation.
The paper tackles the problem of estimating graphons from sparse networks while preserving node-level differential privacy, ensuring that the output hides individual vertex insertions or removals. It provides algorithms with consistency guarantees and explicit error bounds that match or improve on non-private methods under certain conditions.
We design algorithms for fitting a high-dimensional statistical model to a large, sparse network without revealing sensitive information of individual members. Given a sparse input graph $G$, our algorithms output a node-differentially-private nonparametric block model approximation. By node-differentially-private, we mean that our output hides the insertion or removal of a vertex and all its adjacent edges. If $G$ is an instance of the network obtained from a generative nonparametric model defined in terms of a graphon $W$, our model guarantees consistency, in the sense that as the number of vertices tends to infinity, the output of our algorithm converges to $W$ in an appropriate version of the $L_2$ norm. In particular, this means we can estimate the sizes of all multi-way cuts in $G$. Our results hold as long as $W$ is bounded, the average degree of $G$ grows at least like the log of the number of vertices, and the number of blocks goes to infinity at an appropriate rate. We give explicit error bounds in terms of the parameters of the model; in several settings, our bounds improve on or match known nonprivate results.