My research interests span a wide array of topics around graph neural networks. Rather than focusing on a particular niche, I like to constantly look for new interesting and impactful problems. Some of the problems in graph neural networks that have caught my interest so far are:
- Scalability: How do we scale GNNs to billion nodes graphs?
- Dynamic Graphs: How do we learn on graphs that change over time?
- Missing Node Features: How do apply GNNs to graphs where we only observe only a subset of features for each node (which is almost always the case in practice)?
- Low Homophily Graphs: How do we design GNNs that work on graphs with low label homophily (where neighbors tend to have different labels)
Below is a list of my publications in reversed chronological order.