Knowledge Representation and Extraction at Scale – Shubhi Tyagi, Amazon


Session Outline

In this session, the audience will get an overview of the Knowledge Graph (KG) used in Alexa. We will go through three main challenges faced at scaling this technology:

1. Completeness: How to evaluate if the KG is rich in relevant information?

2. Extraction: How to extract facts at scale from unstructured and structured data sources?

3. Verification: How to automatically verify if the extracted facts are true?

Key Takeaways

  • From this session, the audience will get a deeper understanding of how data is represented in a knowledge graph.
  • What are the challenges of scaling a knowledge graph. How those challenges can be addressed at scale.

Add comment