blue circle with dots

Perspectives

Natural language generation and data science

Interview with Kris Hammond, chief scientist, Narrative Science

Deloitte practitioners recently sat down with thought leaders across the spectrum of cognitive computing and data science to discuss current issues and future trends. In this video series, Deloitte’s David Steier, managing director, Deloitte Consulting LLP, talks with Kris Hammond, chief scientist, Narrative Science, about Natural Language Generation (NLG)—what it is, what organizations should be doing with it, and what its future might look like.​

What is natural language generation?

Surrounded by a wealth of data that’s growing all the time, we look to machines to make sense of it. Natural language generation is part of a larger ecosystem in artificial intelligence, cognitive computing, and analytics that helps us turn data into facts and draw important conclusions from those facts. This is how we can make data highly useful and highly relevant in a contextual way. Learn how natural language generation takes facts that machines can understand and turns them into a language that humans can process and act upon.​

What should organizations be doing with natural language generation?

Ready to realize the value of natural language generation? Here are some questions your organization can ask to help you decide:

  1. Do we have the data we want to make sense of? Data has to be available in a systemic way for natural language generation to do its job.
  2. Do we know what story we want to tell with the data? It’s important to understand what narrative the data will inform and what information people are seeking from the data.
  3. What are we going to get from this data? This varies from organization to organization–some businesses achieve greater reliability, others improve quality, while others boost speed and scale.​

What’s next for natural language generation?

With advances in technology like cognitive computing and natural language generation, looking ahead two to five years can reveal—and inspire—what’s possible. As Kris Hammond explains, natural language generation clears two paths to greater understanding. One path follows the rise of chat bots—the ability to interface with services through text and voice conversationally.

Another path for natural language generation involves its integration with other intelligence systems. This application helps machines explain themselves, to tell us not only the results, but how those results were derived.

How far will natural language generation take us, and what are some areas ripe for expansion?

  • Fraud and anti-money laundering: Applying NLG to the narrative component of suspicious activity reports can assist with anti-money laundering compliance. Internal audit teams can also gain new insights into fraudulent activity.
  • Compliance: Natural language generation enables compliance teams to automatically identify the most interesting and important information trapped in structured data and produce language that provides situational context, explanations, and potential next actions.
  • Forensics: With forensic investigations, natural language generation, combined with visual analytics, can reveal insights around anomalous information quickly, which helps target areas for further investigation.​
Did you find this useful?