Natural Language Understanding (NLU) is a form of artificial intelligence that adds some heat to thechatbot fire.
When users engage in a conversation powered by NLU, the results are generally better. You can change your mind about an original request or even interrupt yourself mid-sentence, and you can use unusual words or phrases. The NLU engine is more like a neural network that understands true intent and meaning, and it has a wider vocabulary.
Now, a company called Pat is ready to show the world how users can benefit.
AtMobileBeat 2016, the company announced a seed round of $2.6 million that will help it build an NLU engine.
Pat was founded in 2015 just up the road from HP in Palo Alto. The team is already making an API available in a private beta and hopes to demonstrate how NLU can improve any chatbot.
The developers are using an RRG (Role and Reference Grammar) Model that parses everyday language to discover its “true meaning.” For example, you could tell Siri, “Call Bob, no, I mean Tom,” and Siri would understand that you have changed your mind. Most A.I. tech available today does not parse these exchanges correctly and would get confused.
“To have a meaningful conversation with machines is only possible when we match every word to the correct meaning based on the meanings of the other words in the sentence – just like a 3-year-old does without guesswork,” says Wibe Wagemans, CEO of Pat, in a press release. “Pat has made that radical breakthrough with certainty instead of probability, and can now scale the blue ocean of uncontested A.I. market space in human-machine conversation.”
The company is automating the machine learning that goes into text and voice chats with bots. Pat uses techniques like Word Sense Disambiguation, Context Tracking, Machine Translation, and Word Boundary Identification to accomplish this.
The company says the A.I. field is exploding, and reports bear that out. According to IDC, bots like Microsoft Cortana and Apple Siri will create revenue of around $1.4 billion in 2016 alone. By 2020, the IDC estimates a market for A.I. apps in the neighborhood of $40 billion.