From the course: LinkedIn AI Academy AI-100: 3 Scaling AI at LinkedIn
Join today to access over 24,700 courses taught by industry experts.
Standardization part 2: GNN applications
From the course: LinkedIn AI Academy AI-100: 3 Scaling AI at LinkedIn
Standardization part 2: GNN applications
- With over one billion nodes and over 250 billion edges, the economic graph itself represents an excellent dataset for machine learning. Using techniques like graph neural networks or GNN's we can process all of this massive dataset and learn representations for each of the nodes in the economic graph, and use them to solve important problems across LinkedIn. Many of the most important entity types in the economic graph occur millions of times, and each one of those occurrences can be a training example for graph neural networks. Consider job postings. We have millions of job postings on LinkedIn, and each one will be represented as a node in the economic graph. Each of these nodes in turn, will be connected to a variety of other nodes of different types such as skills, titles, and companies. Collectively, all of these job postings represented as nodes and all of their connections in the graph constitute an excellent…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.