Associate Professor, University of California, Los Angeles
2 papers at NeurIPS 2025
A consistent scalable two-sample test for unlabeled networks based on spectral matching.
Graph convolutions on node features admit a natural central limit theorem which has consequences to multi-class node classification.
