It’s The Same Hamburger!!
NATURAL LANGUAGE PROCESSING (PART I)
The following is part of a series of articles on NLP. (Check Part II & Part III)
Why NLP ?
Human beings are considered as the most advanced species on planet Earth, and what made us manage to achieve outstanding success in the first place is our ability to communicate and share information and experiences.
That’s where the concept of developing a language comes in, which becomes one of the most diverse and complex part of our existence, with over 6500 different languages around the world.
How Can Computers manage to Understand Human Language?
According to the industry estimates, only 20% of the available data is present in the structured form, all the remaining 80% represent data under unstructured form; it’s whether data being generated as we speak, as we post on Twitter, as we send voice mails on WhatsApp or we text messages on Messenger. Today, majority of the data exist in the textual form, which is highly unstructured, that’s why it’s important to get familiar with text analysis techniques and natural language processing, in order to produce significant and actionable insights from this data.
What is NLP?
NLP in a nutshell is the process of deriving meaningful information from natural language text it usually involves the process of structuring the input text deriving patterns within the structured data and finally evaluating and interpreting the output.
Industry use cases of NLP
I’d be lying to you if I tell you that computers could understand human language the same way we do, but what I can assure you that they already can do a loooot! And guess what? NLP is the main reason behind all that.
Below, I handpicked some examples of cross-industry use cases of NLP for several business purposes:
• Virtual assistant:
It’s basically an intelligent system capable of understanding human voice without the requirement for keyboard input. Based only on voice input, virtual assistance devices can perform several tasks as organizing calendars, sending text messages, managing emails and do lists. In other word, it’s kind of a personal smart assistant that you hire for free (well not really for free.. let’s say you just pay him a signing bonus), who provides a 24/7 service, doesn’t take breaks or holidays, smart enough to speak and understand a bunch of languages and available all the time via your SMART phone, SMART watch, SMART speaker… just to make your life easier, SMARTly.
You can experience this great functionality with Cortana developed by Microsoft, Siri by Apple or Alexa by Amazon …
• Chatbots:
As the name implies, it’s a bot designed to conduct conversations with human users. Chatbots took the customer support experience to a new level by saving time, energy and human resources. Before chatbots came in, the customer support experience looked pretty much like this : You call a “hotline” number, you expect 10mn waiting on the line (and you get charged for it), you listen to the most boring song ever while waiting … If you survive these 10mn, sometimes even more, and you’re lucky enough they didn’t cancel on you, you finally get to speak to “the expected one”.. it turns out it’s a man/woman sitting in a loud noisy room, that sounds like a Wall Street trading room, you barely could hear him/her, the brouhaha drives you crazy and the call end up with the famous saying “we took note of your request and we will be in touch with you soon” .. never happens. Now, with chatbots, it’s a completely different experience. It’s a free service, available 24/7 and smart enough to answer all your questions. In case it couldn’t answer your request, it redirects you to the person in charge to serve you better. More than that, it saves your conversations history to follow up with you and provide you with a customized experience. It’s pretty much as you hired a personal consultant to handle any kind of issue you may have with a product or a service. This way, we don’t only prevent bombarding the support team with useless phone calls from annoying customers, but we also have the possibility to access information quickly and efficiently in order to increase productivity.
• Neural machine translation:
What has previously seemed like a foolish attempt to replace, even imitate the professional translation has now incredibly improved. Thanks to NLP, today some Artificial Intelligence technologies are capable of translating tons of text, in any language, just in a twinkling. Neural Machine Translation (NMT) has taken this experience even further by providing large artificial neural network models, trained jointly (end-to-end), able to predict a a sequence of words likelihood, in order to maximize translation performances.
I know it rough to admit it but NMT is behind job cuts of hundreds of professional translators today, these models are performant enough that they can simulate, even bypass, professionals work. Well that’s what technology does, changing jobs continuously with new ones so we can evolve as Human beings.
In 2016, Microsoft pioneered the launch of neural machine translation technology with Microsoft Bing, paving the way to Google and Amazon who are trying to deliver sophisticated machine translation tools on the market.
• Sentiment Analysis:
In the purpose of brand monitoring and adjusting sales strategy, it’s essential to analyses customer feedbacks and figure out their expectations.
This is where sentiment analysis comes in as the process of computationally identifying and categorizing opinions expressed in a piece of text, retrieve significant insights to determine whether the writer’s attitude towards a topic, product, etc. is positive, negative, or neutral.
Through sentiment analysis technologies, companies can analyze customers feedbacks through blogs, comments or social media, and identify what do they think about their products or services.
Microsoft Azure provides a stack of cognitive services, which are intelligent algorithms that “bring the agility and innovation of cloud computing to your on-premises environment”. Azure Cognitive Services provide a powerful API: Text Analytics API, which allow users to integrate pretrained NLP algorithms to their code. Text Analytics API includes a Sentiment Analysis capability “useful for detecting positive and negative sentiment in social media, customer reviews, and discussion forums”, to enhance the Customer Relationship Management.
