In this vertical we unleash the power of artificial intelligence to analyze complex graph structures, enabling insights and advancements in diverse fields such as social network analysis, cybersecurity, and recommendation systems. In this regard a survy has also been conducted for graph data augmentation.
● Data Augmentation for Graph Data: Recent Advancements, arXiv, 2022
a. Higher Order Sparse Convolutions in Graph Neural Networks
Graph Neural Networks (GNNs) have been applied to many problems in
computer science. Capturing higher-order relationships between nodes is crucial
to increasing the expressive power of GNNs. We introduce new higher-order
sparse convolution methods based on the Sobolev norm of graph signals.
Related Publications
b Reconstruction of Time-Varying Graph Signals
Graph Signal Processing (GSP) extends digital signal processing to graphs and
finds applications in sensor networks, machine learning, and image processing.
This work focuses on time-varying graph signals, introducing the new algorithms
based on Sobolev smoothness for signal reconstruction.
Related Publications
c. Network Community Detection
This study tackles community detection challenges in network structures by
leveraging the geodesic space. Overcoming the limitations of traditional graph
theoretic algorithms, we explore the methods that employ sparse linear coding
with L1 norm constraint and sparse spectral clustering algorithms to enhance
accuracy in identifying community boundaries.
Related Publications