Alejandro Ribeiro

DiPhon enables the generation of large graphs from small training samples while preserving their core topological properties, revolutionizing scalable graph generation.

Linear attention fails to capture spectral variations in graphs, but Graph Convolutional Attention achieves superior denoising by directly utilizing the graph spectrum.

Ditch iterative optimization for wireless resource allocation: a single sample from a learned graph diffusion model achieves near-optimal performance with strong generalization.

Graph Transformers can be effectively trained on small graphs and then transferred to larger graphs, thanks to their GNN-based positional encodings, offering a path to scalable graph learning.