Graph Signal Processing

Graph signal processing (GSP) analyzes data residing on graphs, aiming to leverage the underlying network structure for improved signal processing tasks. Current research emphasizes online and adaptive methods for handling dynamic graphs and incomplete data, employing techniques like graph filters, neural networks (including graph convolutional networks and transformers), and Kalman filtering for signal estimation, reconstruction, and denoising. These advancements are impacting diverse fields, including recommender systems, brain imaging analysis, and network inference, by enabling more efficient and accurate processing of complex, interconnected data.