Building and deploying knowledge graphs

The Data Exchange Podcast: Mayank Kejriwal on the critical role knowledge graphs play in modern AI applications.

Subscribe: Apple • Android • Spotify • Stitcher • Google • RSS.

In this episode of the Data Exchange I speak with Mayank Kejriwal, a Research Assistant Professor in the Department of Industrial and Systems Engineering, and a Research Lead at the USC Information Sciences Institute. The focus of our conversation is knowledge graphs, a collection of linked entities (objects, events, concepts). While there are a variety of ways knowledge graphs are described, a 2016 paper proposed the following succinct framing:

A knowledge graph acquires and integrates information into an ontology and applies a reasoner to derive new knowledge.

Knowledge graphs are used in many AI applications. For example, Google uses a knowledge graph to enhance its search engine results with infoboxes that appear in some search results. Other areas where knowledge graphs are common include e-commerce, healthcare, and financial services.

Mayank is passionate about knowledge graphs and has used them for a variety of applications in media, the life sciences, public policy (human trafficking), and more. He is working on a new book aimed at practitioners (Knowledge Graphs: Fundamentals, Techniques, and Applications”).

One of things we discussed was the use of graph databases for applications involving knowledge graphs. Mayank believes that for many use cases, tools used for search applications or relational databases, are better suited for storing knowledge graphs:

“For the graph database people, if they can prove to me and everyone else that they can run faster than a relational database in the general case. Or they can run faster than Lucene for information retrieval. Then I will use graph databases. I don’t think they are. … Speed is very important. Google is not using a graph database to our knowledge.”

Subscribe to our Newsletter:
We also publish a popular newsletter where we share highlights from recent episodes, trends in AI / machine learning / data, and a collection of recommendations.

Related content:

• A video version of this conversation is available in our YouTube channel.
• Denise Gosnell: “How graph technologies are being used to solve complex business problems”
• Edo Liberty: “How deep learning is being used in search and information retrieval”
• Matthew Honnibal: “Building open source developer tools for language applications”
• Alan Nichol: “Best practices for building conversational AI applications”
• Amy Heineike: “Machines for unlocking the deluge of COVID-19 papers, articles, and conversations”
• Edmon Begoli: “Hyperscaling natural language processing”

[Image by anncapictures from Pixabay]