By: Bernard Marr
Originally published at data-informed.com
Do you talk to your computer or smartphone? Just a few years ago, that question would have been absurd. But with advances in natural language processing, the likelihood is that you have asked your phone to send a text or search the web for something within the last day.
In fact, natural language processing (NLP) is one aspect of machine learning, big data, and artificial intelligence that has the potential to truly change everything.
In its most basic terms, natural language processing is the ability of a computer to understand natural human speech as it is spoken. It’s the difference between saying, “Siri, where’s the nearest coffee shop?” and, “Search coffee shops ZIP Code 80021.”
For a long time, searches online had to be done by typing in strings of words combined with Boolean search terms that ended up looking and sounding nothing like a conversation. Now, however, you can type a question into Google exactly how you’d ask it to a friend, and Google can reliably provide a good answer.
The same recognition of natural language is being developed for speech. AI assistants like Siri, Cortana, and Google Now are good examples of this.
While it seems simple for a human to answer a natural language question, it’s an incredibly complex task for a computer, requiring many steps computations and predictions, all of which must happen in the cloud and in a split-second.
The fascinating thing is that, while a human inherently understands what is being said, a computer cannot really be said to understand language. It can parse out the different words, the context, the grammatical usage, etc. and then make a prediction about which response will be the best, but it does not actually understand what we are saying.
One goal of NLP is to do away with computer programming languages like Java, Ruby, or C and replace them with natural human instructions and speech. Another ultimate goal is realistic artificial intelligence, wherein the computer can react to and interact with a human flawlessly.
Computer “assistants” like Siri and Cortana are the most visible use of NLP today, but there are many other applications of NLP in use. As mentioned above, Google has poured a great deal of resources into NLP as it relates to search, allowing us to type or speak a natural question and receive a relevant answer. Google also is using NLP to create predictive text responses to emails in its Inbox email client, allowing users to choose from one of three responses and respond to an email with a single click.
You may have used NLP for yourself if you have ever used the “translate” link inside Facebook to translate a foreign language into your own (with varying results) or used Google translate on Google or Bing search results. A reliable machine translation has been a goal of NLP since the 1950s, and results are improving all the time.
Other programs are being developed and used that can automatically summarize long documents or extract relevant keywords for searching. The legal system is using these types of applications, for example, to help lawyers sort through thousands of pages of documents in any given legal case to find relevant information.
Marketers are using NLP for sentiment analysis, combing the millions of tweets and other social media messages to determine how users feel about a particular product or service. It has the potential to turn all of Twitter or Facebook into one giant focus group, at a fraction of the cost.
Another way you likely use NLP daily in your life is with text classification – which is what Google and other email providers use to determine if an email is spam or not. This is a very simple binary classification: an email either is spam or it isn’t. But more sophisticated forms are being used for such complex analyses as determining the author of a work by comparing it to other works.
Companies are predicting that chatbots will be able to take over some customer-service functions in as little as five years, providing automated, real-time responses to simple customer-service problems and questions.
Integrations also are being developed for particular situations and users. For example, one company has developed an interface for the Amazon Echo that can allow business leaders to track key performance metrics. In fact, when the system is set up, a colored light bulb in an office can be used to visualize those metrics. One user set the system to monitor hold times for customer service, and when the light bulb goes red, he knows there is a problem that needs to be addressed immediately.
How NLP will Change Things in the Future
Bernard Marr is a bestselling author, keynote speaker, strategic performance consultant, and analytics, KPI, and big data guru. In addition, he is a member of the Data Informed Board of Advisers. He helps companies to better manage, measure, report, and analyze performance. His leading-edge work with major companies, organizations, and governments across the globe makes him an acclaimed and award-winning keynote speaker, researcher, consultant, and teacher.