Graph Neural Networks preserving categorical features in Embedding Representation

What is it about?

Graph Neural Networks (GNN) can be utilised to solve various tasks like peer-set generation, recommendation, community detection, and anomaly detection. However, most of the real-world data come with high cardinality categorical features such as merchant industry and current GNN algorithms are not designed to work for such sparse features.

Featured Image

Photo by Conny Schneider on Unsplash

Photo by Conny Schneider on Unsplash

Why is it important?

The research addresses the above problem using novel framework CaPE (Category Preserving Embeddings) which generates category-preserving embeddings using two GNN modules trained sequentially under the different settings of loss functions, input features and attention mechanism.