We’re seeing the commercial use of artificial intelligence (AI) more and more with the introduction of devices like Alexa and Google Home. People use these tools to complete all sorts of tasks such as verbally place orders for goods and services or ask for directions to a place they're unfamiliar with.
AI also continues to transform the way companies can provide people with a better customer experience. AI is now using Natural Language Understanding (NLU) applications with IBM’s Watson to help medical providers collect and interpret clinical data from patients – and that’s just one example.
But how does it all of this technology come together? What is it that allows you to issue a command and get back what you need from an AI enabled application?
Helping Machines Understand Us
Machines use their own language for processing commands and performing tasks. In order to directly interact with people, they need a cognitive way to turn human language into a syntax that the computer can read. That’s where Natural Language Processing (NLP) comes in.
Let’s say you chat with Alexa and issue a voice command. NLP takes your words and parses them into machine language. You then get a response within seconds that answers your question.
Is that it? Not even close. At this point Alexa has the words from your query, but has no idea what you’re asking for or why. Think of NLP as your smartphone. From the outside, you see the sleek design that presents visual and audio confirmation of your requests. But inside there’s a lot more going on.
Helping Machines Process Our Intent
If NLP is your new smartphone, NLU is the engine inside,working hard to keep things going. NLP captures your command and turns it into machine language. But your automation still has no idea what you actually want.
NLU takes that machine language and uses various algorithms to parse out your actual intent. Let’s say you’re in a chat with Alexa. You ask, “Alexa, give me directions to New Orleans”. NLU breaks it down the following way:
Directions [intent] New Orleans [location]
Once NLU understands that you need directions to New Orleans, it can take things further. It knows it needs to gather your location information from your device and do a search for sources that provide it with the cleanest match to your query.
This is where NLP kicks in again. NLP still has no idea what you want or why you asked for it, but it knows that it needs to be put to you in a way you understand. It takes that information and turns it into a verbal or visual response to presents it back to you in human language.
Continuing To Learn and Grow
NLU is the heart of any AI system. NLP facilitates an interactive voice response between a human and machines and allows NLU to receive the information needed to interpret and respond to human commands.
The continuing challenge for computer scientists and programmers is designing NLU algorithms that adequately interpret and understand the variations in human language. To do that, AI must continuously interact with humans so it can learn and grow as our language does.