“People have always communicated through stories and language, why should we expect them to change now?”
- Harvard Business Review
Here are answers to the top five questions regarding natural language generation (NLG).
1) What is Natural Language Generation?
Natural Language Generation (NLG), a subfield of artificial intelligence (AI) which produces language as output on the basis of data input, is not a new concept. What is new, however, is the increase in adoption of NLG into the enterprise. There are a plethora of ways the technology is being employed, primarily to improve human productivity, customer engagement and operational efficiency.
2) What’s the goal of NLG?
People have always communicated ideas from data. But with the explosion of data that needs to be analyzed and interpreted, coupled with increasing pressures to reduce costs and meet customer demands, the enterprise must find innovative ways to keep up.
As it turns out, a machine can communicate ideas from data at extraordinary scale and accuracy. And it can do it in a particularly articulate manner. When a machine automates the more routine analysis and communication tasks, productivity increases and employees can focus on more high-value activities.
As stated in the book, "Practical Artificial Intelligence for Dummies":
“The goal of natural language generation (NLG) systems is to figure out how to best communicate what a system knows. The trick is figuring out exactly what the system is to say and how it should say it. Unlike NLU (Natural Language Understanding), NLG systems start with a well‐controlled and unambiguous picture of the world rather than arbitrary pieces of text.
Simple NLG systems can take the ideas they are given and transform them into language. This is what Siri and her sisters use to produce limited responses. The simple mapping of ideas to sentences is adequate for these environments.”
3) How is NLG different than NLP?
Gartner’s recent Hype Cycle for BI and Analytics sums up the difference between NLG and NLP (Natural Language Processing) well:
“Whereas NLP is focused on deriving analytic insights from textual data, NLG is used to synthesize textual content by combining analytic output with contextualized narratives.”
In other words, NLP reads while NLG writes. NLP systems look at language and figure out what ideas are being communicated. NLG systems start with a set of ideas locked in data and turn them into language that, in turn, communicates them.
4) What are the different variations of NLG?
Basic NLG: Basic NLG automatically translates data into text via Excel-like functions. An example of this would be a mail merge that restates numbers into language.
Templated NLG: Here, the user is responsible for writing templates, determining how to join ideas and interpreting the output. Essentially sentence building, it relies on business rules, basic calculations (ex: sum) and templates with boilerplate text to automate content. Templated systems are limited in their ability to draw from multiple data sources, perform advanced analytics, achieve reusability from one project to the next and explain how it came to the story it created, with no understanding of what the user is trying to communicate or their particular domain.
Advanced NLG: Advanced NLG communicates the way humans do – infusing intelligence and intent into the process from the very beginning. It assesses the data to identify what is important and interesting to a specific audience, then automatically transforms those insights into Intelligent Narratives – insightful communications packed with audience-relevant information, written in conversational language. Backed by a knowledge base, Advanced NLG systems understand the domain and can write contextually about a user’s business at a scale that is not possible by humans.
5) What is the future of NLG?
Alexa, Cortana and others are ushering in the era of intelligent personal assistants, helping to make everyday tasks easy and efficient for consumers. The enterprise is catching up, with conversational interfaces that are facilitating engagement across employees and to customers, raising the bar on how these systems communicate.
What is the critical differentiator from a conversational bot that performs tasks to one that engages, explains, and illuminates? Advanced NLG.
Per a recent article in the Harvard Business Review, "Bots that Can Talk Will Help Us Get More Value from Analytics":
"Conversations with systems that have access to data about our world will allow us to understand the status of our jobs, our businesses, our health, our homes, our families, our devices, and our neighborhoods — all through the power of Advanced NLG. It will be the difference between getting a report and having a conversation. The information is the same but the interaction will be more natural."