Skip to main content

Natural language processing bridges the gap between computers and humans

(Image credit: Image Credit: John Williams RUS / Shutterstock)

Between Siri, Cortana, and Alexa, artificial intelligence has hit the mainstream. While people converse with AI on a regular basis, the average person knows far less about the natural language processing that makes the technology possible. 

Many people have observed NLP in action without realizing it. These AI helpers can listen to user commands and turn around to answer questions, play music, schedule meetings, or even control other connected devices. Suddenly, J.A.R.V.I.S. from “Iron Man” doesn’t seem so far-fetched. 

NLP isn’t limited to verbal exchanges, though. The technology also allows computers to quickly process written words to perform countless complicated tasks. Sophisticated chatbots can field complaints, troubleshoot technical issues, and guide customers through online retail. The absurd utility of NLP powers many of the most popular apps and devices in the world. 

This advancement has changed the computer-human dynamic in fascinating ways. At a foundational level, it has made it possible for computers and humans to speak the same language. That seemingly minor breakthrough has opened the floodgates for innovation. 

Speaking My Language 

Computers were once seen as “dumb terminals,” and corresponding human interactions followed a “garbage in, garbage out” mentality. The only language computers spoke consisted of an extensive, cryptic set of hand-coded, programmed rules. 

Application program interfaces allow developers to quickly and easily incorporate NLP functionality into their creations. Some of these APIs are simple and focus on a single task, such as seeking out and extracting content or rating a specific input’s emotion to gauge favorability. Others, like IBM Watson, allow integration with more complex cognitive capabilities. 

H&R Block uses IBM Watson to train its online tax service on more than 74,000 pages of tax code. Consumers inevitably have plenty of questions and concerns as they file their taxes, but IBM Watson has developed the smarts to suggest tax credits and deductions based on how clients respond to questions. Users trust the functionality to provide the correct tax information, and the software is able to interpret complex data quickly and efficiently. 

NLP-enabled apps act on data instead of only viewing it. At the same time, the rise of real-time data has added fuel to the NLP fire. Computers can now process language on the fly, crunching data as it rolls in. 

Considering the vast — and perpetually growing — amount of data in the world, this real-time processing is good news. APIs and services have a wealth of information to access, improving NLP’s intelligence-gathering capabilities as a result. NLP can effortlessly tap into and sift through this data, which would take humans far longer to process manually. 

As technology evolves and processing power increases, computers will be able to use NLP to synthesize more data faster. This confluence of changes enables NLP to rapidly crunch bigger sets of more robust data while accurately analyzing complex human inputs. 

Music to Our Ears 

Devices such as the Amazon Echo and Google Home are everyday examples of products that use NLP technology to respond to human commands. As evidenced in the earlier example of H&R Block, next-level applications have demonstrated the technology’s incredible potential. 

Alex da Kid, a British music producer, recently collaborated with IBM Watson to create the song “Not Easy.” The creative process began with an API that culled through five years of text — including Supreme Court rulings, news articles, movie synopses, and Wikipedia articles — to find common cultural themes and combined that knowledge with emotions people expressed on social media channels. A different API then analyzed the lyrics and genres of more than 26,000 popular songs. 

With that data in hand, Alex da Kid worked with Watson to craft a song framed around the emotional theme of heartbreak. Artists can actually have Watson listen to a piece of music and scan it for trends, then ask the AI to offer a version with different elements —whether it's a reggae beat or dance hall sound. 

“I love the progression of technology,” Alex da Kid told Rolling Stone. “This [collaboration] is the absolute embodiment of that. This wouldn't have existed without Watson.” 

Listening to the song, it’s not obvious that a computer played such a pivotal role in the creative process. Clearly, NLP has entered the mainstream. Whether it’s chatbots helping you prepare your taxes or AI influence in the latest music, these endeavors are pushing NLP to the next level. 

Why NLP Is Here to Stay 

Technologies naturally come and go. Remember when Adobe Flash and CD burners were all the rage? Many creations burst onto the scene but fade into the background after the initial attraction wears off. 

NLP isn’t an overnight sensation — it’s the culmination of research going back to the post-war research boom of the 1950s. An early experiment saw machines translate Russian sentences into English. After decades of slow progress, NLP has flourished because statistical methods can finally be applied to massive amounts of text. 

Computing power becomes cheaper with each passing day, and society’s shift to digital means everything is in a format that NLP can readily analyze and process. As NLP becomes cheaper and easier to use, we’re certain to discover new and innovative ways to capitalize on the service. 

The same digital communication that has fed the rise of NLP also provides a use case for the technology. Online communities need moderators to deal with trolls, and NLP can help companies scale moderation without breaking the bank. Likewise, NLP enables companies research their customers’ online habits to better target marketing campaigns. 

Above all, NLP allows people to communicate with computers in a natural way. Humans are genetically built for spoken language — it’s our most natural communication method — and modern algorithms for speech recognition and synthesis have grown to the point that they can be embedded in cheap consumer devices. 

Used correctly, NLP can bridge the gap between the digital world and the human voice. By continually developing NLP’s functionality, we can effectively train computers to speak our language. 

Rogue AI in films such as “Ex Machina” and “2001: A Space Odyssey” might make us a bit leery of that concept, but NLP will be the tool that propels the next wave of breakthrough technologies. All we have to do is start talking. 

Image Credit: John Williams RUS / Shutterstock

Josh Marinacci
Josh Marinacci is head of developer relations at PubNub, a data stream network that enables customers to build and manage real-time functionality for web and mobile applications and IoT devices.