Eric MagangaDecember 12, 2019

Natural Language Processing: what you need to know for 2020


You might already benefit from Natural Language Processing whether or not you know it. The technology behind systems we use every day such as voice assistants, auto-correct and spell check is prime to continue to play a large role in our lives. Get to know the most important features of this technology in this blog.


Natural Language Processing (NLP) is the use of Machine Learning to understand and analyze natural speech. It combines elements of linguistics with elements of computer science. It has different components that are used for understanding and gaining context that we will get into later, but let’s start with why even bother with Natural Language Processing in the first place?

Why Natural Language Processing is important

SignAll a Natural Language Processing case study

The Hungarian startup SignAll uses Natural Language Processing and computer vision to transcribe sign language.

The company uses an XBOX Console’s Kinect camera and three other RGB (red, green, blue) cameras to track follow and monitor hand shapes, facial expressions and body movements in order to translate these movements into written speech.

Hopefully they will inspire more people to use this technology for good causes and to hopefully improve people’s quality of life. Natural Language Processing applications for sales and marketing

At Ocean, we use Natural Language Processing to interpret information about similar companies so you can get them organized in neat clusters to use for your marketing and sales efforts. With the information gained from scanning the websites you want to know about, you have solid data that you can use to go forward with your business without working off of assumptions.

Corti: using NLP to combat heart attacks

Copenhagen company Corti, have used Natural Language Processing as their foundation to grow their business beyond Denmark. By evaluating conversation patterns, Corti can let medical professionals know about signs that you may be having a heart attack.

Saving lives with a search engine

Baidu, the main search engine used in China (90% of all searches made in mainland China are made on this platform), uses a chatbot where patients can inform their doctors about their symptoms in a streamlined process. Therefore it can improve the medical care experience for many people. Natural Language Processing is also generally used in several chatbots to understand queries.

But where is NLP headed next?

Let’s explore:

Natural Language Processing trends for 2020

There are some technological trends that use Natural Language Processing to great effect meaning it is most likely already part of your life.

For example voice assistants are built around this technology. Therefore you probably already have Natural Language Processing as a part of your smartphone and/or smart speaker setup. The increasingly popular Google Home and Alexa devices employ Natural Language Processing to interpret your voice and answer your queries and questions.

SEO TIP: To optimize your written content for these devices write content that features both the question and the answer to a query because these are easiest for these devices to understand.

For example for a query about Michael Jordan’s height it would be best practice to have content that said: “How tall is Michael Jordan? He is 6’6”. This can for example be done with an FAQ page.

Key takeaway: Natural Language Processing will continue to be important because it can improve the quality and convenience of people’s lives.

Natural Language Processing Examples

Spam filters use Natural Language Processing to make sure when phrases like “buy now”, “completely free” and “you have won a million dollars” are used in emails, these messages are quickly sent to your spam folder.

Auto-correct which is used when we write messages on our phones, utilizes Natural Language Processing to predict what you mean to write based on context.


Natural Language Processing uses syntax to make sense of words grammatically.

Sentiment analysis is similar to the concept of social monitoring but on a smaller scale. It consists of: finding out if opinions are negative, positive or neutral. For example your Facebook page reviews can express your audience’s, criticism, praise or indifference for your products or services. Sentiment analysis can also be used for figuring out your audience’s emotions (happy, sad, angry) and intentions (likely or unlikely to purchase).

Semantic analysis is based on history and context. It goes beyond sentiment and deals with the meaning and relationships between words as well as how they are positioned together. Semantic analysis can be used to distinguish between joy (happiness) and Joy (a woman’s name).

Uses and applications

Named Entity Recognition (NER) can reveal categories in a complete written text including identifying the main objects, people and locations by simply scanning its contents.

Polysemous words have related meanings. This is a challenge Natural Language Processing faces because these types of words are more complex to analyze. For example park can mean to park a car or a nature park.

Parts of Speech (POS) tagging can break down sentences and tell you which words are nouns and verbs for example. POS tagging can also be used to distinguish between situations where rose is a noun (the red rose) or a verb (Derrick rose up). This can make the difference between a correct and incorrect interpretation of a transcription.

Closely related: what’s the difference between NLP and Natural Language Understanding?

Natural Language Understanding (NLU) is as the name indicates, focused on understanding and comprehension. It is one part of Natural Language Processing along with Natural Language Generation (NLG) which turns data into cohesive language. The advantage of this is that it executes the process much faster than a person could. It is Natural Language Generation that determines the semantics behind words.

Text mining is understanding facts, relationships and claims in text. After the meaning is gathered, it is structured and visualized for example as mind maps or tables. Text mining transforms text to data. It can for example be used in tickets/queries on a website for determining the urgency of a ticket. By identifying certain words in the query it will automatically group it as urgent. It is also how Ocean gets you your list of similar companies.

Bonus section #1: Learning about NLP (next steps)

Course providers like Udemy and Coursera offer a chance to learn even more about Natural Language Processing. So if your curiosity is sufficiently piqued, head over to one of these sites and get started for a chance to understand even more of the intricacies of Natural Language Processing.

Bonus section #2: All the buzz about BERT

The Google update simply known as BERT (Bi-directional Encoder Representations from Transformers) is an example of Natural Language Processing that has been on the minds of marketers and Search Engine Optimizers lately.

When Google announced they were implementing it as the biggest algorithm update since their rankbrain update, it really came to prominence. It started out impacting one in ten English searches, improving the comprehension in otherwise ambiguous sentences that Google struggled to understand in the past.

Bing searches have used it since earlier this year and for more than only a limited number of searches but when Google embraced it, its notoriety grew a lot more.

It has since been rolled out in over 70 languages where Google representatives have also confirmed it will continue to impact one in ten searches.

For example for the search “do you need a visa to travel from Brazil to the USA”, it will correctly interpret that you are travelling from Brazil to the USA and not the other way around and will direct you to the US embassy website’s visa section.

Bonus section #3: Funny applications

Natural language processing has limitations when it comes to multiple languages being used at once. In The Simpsons, Bart’s nemesis Sideshow Bob has “Die Bart Die” written on him and argues that it is German “The Bart The”. This is the reason Twitter categorized a German user writing “Die Boomer” as hate speech and banned their account for 12 hours.


Natural Language Processing might already be tied into your life in ways you haven’t realized and for good reason as it has the potential to make people’s lives convenient and of a higher quality. In some cases with medical care it can go as far as to save lives. So if you are for example getting help for heart trouble in a place with NLP support for this, you will be one of many benefitting from this technology.

In the mood for sharing?