<ul><li>Graphons are important in statistical analysis of network data but existing estimation methods struggle with scalability and resolution-independent approximation.</li><li>A novel, scalable graphon estimator using moment matching and implicit neural representations is proposed in this work.</li><li>The estimator avoids latent variable modeling and leverages empirical subgraph counts for direct estimation, providing a polynomial-time solution.</li><li>The proposed method achieves strong empirical performance, demonstrating high accuracy on small graphs and superior computational efficiency on large graphs compared to existing scalable estimators.</li></ul>

A Few Moments Please: Scalable Graphon Learning via Moment Matching

Discover more