PhD student, University of Oxford
1 paper at NeurIPS 2025
We formalize the over-squashing phenomenon in spatiotemporal graph neural networks and analyze how it affects information propagation across the spatial and temporal dimensions.
