Author: Caroline Smart
The trouble with NLG is that even after you find out what the three letters stand for (Natural Language Generation), you may be none the wiser. It’s when you understand what NLG does, that you appreciate the world of possibilities it opens up. And that’s particularly true as we step off the Information Super Highway (just some of you may be old enough to recognize the reference) onto the Internet of Things.
There is a lot of artificial intelligence out there. But if you have to choose between a date with your robot vacuum cleaner or R2-D2, which would it be? Okay, if you chose the vacuum cleaner, that would be sad. Obviously R2-D2 wins hands down. And the reason is obvious: the little droid can speak to you. That’s the magic of NLG. It lets the machines tell you what is happening in their world.
So, what are the five things you need to know about NLG?
- NLG software can pick out what is significant in data and tell you the background story. For example, that spike in sales of Stevie Wonder downloads is due to the fact that his Car Pool Karaoke was featured on the Late Late Show again, and after it first aired, his greatest hits album went to number one, so we can expect a similar/repeat performance this time. So, NLG can join the dots by analyzing what else is happening in the wider context, taking into account information in the data – such as current and historical sales – and complementing this by pulling in facts external to the data – such as news sources – to build a correlation to inform the output report.
- It can speak your language. If you are in marketing, NLG can replicate your tone of voice and produce reports just like the ones you would write (if only you had the time), or your assistant would write (if only you had one). And your NLG assistant can spell and knows its grammar. Even if you’re not in marketing, and you just appreciate well-written, tailored reports (i.e. reports that don’t sound robotic), then NLG is your friend – after all, that’s what the ‘N’ for Natural means.
- NLG is versatile: the same data, different messages. The CEO might want to know what’s happening. The CIO might want to know why it’s happening. The Operations Manager might want to know how it’s happening. NLG can tailor reports to suit what each of its users need to know. To the techy audience, it can match their technical language. To the business audience, it can use corporate terms. To the finance audience, it can speak in figures and percentages.
- It’s not all words. NLG can combine data visualization and the supporting narratives to explain them – because sometimes a graph is the best way to display data. Like a human, our NLG technology instinctively chooses the best communication method and combines this in the output report, just like a human expert.
- NLG matures with age. New layers of language are added all the time. Arria’s NLG platform possesses a solid rule-based foundation. This can underpin and combine with other branches of AI – such as Natural Language Understanding, machine learning, predictive analytics, and prescriptive analytics.
Arria’s NLG Platform embodies over 30 years of expertise and experience in the field of computational linguistics, and was built by world experts in this field (two of whom, Ehud Reiter and Robert Dale, have written the go-to book on this subject, Building Natural Language Generation Systems). The Arria Platform is mature, reliable, scalable and can speak like an expert in many sectors. It can alter its tone and style depending on who it is addressing, even when the narratives are based on exactly the same data.
Language might be the most common way for us to communicate, but it is also fiendishly difficult to recreate artificially. If you want your data to find its voice, come and check us out at Arria.
Two ways you can discover NLG today:
- Build your own simple NLG reports using Articulator Lite – our web-based NLG toolkit, currently in beta.
- Request a demo to see our NLG Platform in action and speak to our data scientists to learn how NLG can transform your data.
Author: Caroline Smart, Copy Editor at Arria NLG