Joint personalized search and recommendation with hypergraph convolutional networks

This invention describes a method for training an artificial intelligence system, specifically a Hypergraph Convolutional Network (HGCN), to improve personalized search and recommendation. It works by building a complex network (a hypergraph) that connects users, search queries, and content items. The system then learns patterns from how users interact with search and recommendations, using these patterns to make better suggestions and search results. The claims narrow this to a specific method of *training* such a network.