NLP vs NLU vs NLG: What’s With The NL?
Welcome back to our little corner of the internet!
Today we’re going to be looking into one of the topics that was briefly touched upon in my previous blog post — Natural Language Processing, Understanding and Generation (NLP, NLU, NLG). But what are all of those things — and what’s with the clear running theme Natural Language? Let’s go ahead and learn something new!
So what is Natural Language Processing?
Natural Language Processing is “is a machine learning technology that gives computers the ability to interpret, manipulate, and comprehend human language” according to AWS. The ability to process human language is of course essential in things like Conversational AI, and another good real-life example of a use of NLP is a chatbot. Botpress described NLP as “what makes a chatbot feel human” — and they’re right in saying this, because the ability to comprehend human language allows chatbots to communicate with us in a way that we can understand. NLP is often known as the “link” between human and machine language.
What’s Natural Language Understanding then?
Both NLU and NLG are declared “subtopics” of Natural Language Processing. Natural Language Understanding is pretty self-explanatory, as it is the part of NLP that understands (and analyses) input data from the human; and in turn, it makes sense of everyday human sentences. It works through algorithms that turns speech from humans into a data model of different definitions; so in essence, it breaks down the building blocks and elements of speech!
What’s Natural Language Generation then?
As I said before, NLU and NLG are subdivisions of NLP, meaning they make up two parts of it. NLG (in basic terms) does the other “half of the work” by generating sentences and responses to user commands, which it has understood thanks to NLU. This might be Siri’s response to you asking when your next alarm is, or asking for the time (but obviously many other things happen amidst the NLU and NLG for this to function).
I’ve created a little graphic to aid your understanding of NLP (it’s a black png, so sorry users of dark mode!) — hopefully it helps you visualise how everything works in a simple way.
Why is it important for AI to Understand Human Language?
Well, it makes it easier to integrate AI into our everyday lives without having to know code. For example, imagine telling Siri to switch off your alarm — but not in human language, in C++ (or worse, binary)!
What Has NLP Been Used For In the Past?
NLP has been used for a wide variety of things — some which you may not expect! Let’s look into 3 unique uses of NLP (aside from being in Voice Assistants).
#1 — Email Categories & Filtering
Ever seen the primary, social or promotions filter in your gmail? Well this filtering technique’s backbone is actually NLP — which it cleverly leverages to comprehend the contents of a user’s emails and tag into one of the categories. This may be one of the features you didn’t know about, but still found useful in the long term (or at least you think you did, looking back). This also allows you to have access to the important emails without being distracted by a whole load of spam content.
#2 — Autocorrect
Yup — it’s autocorrect — the, (as per the namesake) auto-corrective technology which attempts to correct any unnoticed spelling errors. It’s one of those things most people either love or hate. Anyways, NLP plays a significant role in the tech behind how it works — it uses its ever-adapting knowledge (it learns the more you use it) of the English language to rephrase your sentences or attempt to find words that you’ve mispelled. The same goes for autocomplete, which allows you to click “the middle suggestion” on your keyboard and still get a somewhat understandable sentence.
#3 — Language Translation Tech
Have you ever used Google Translate but then been told that the translation was incredibly….wonky? Well, worry not, because translation applications used to be even worse — overlooking simple facts (like other languages using different sentence structures). A lot of translating tech today uses NLP to provide more accurate translations and some are even able of detecting the language of text just from the text provided.
Thank you so much for reading this article! If you enjoyed it please do remember to leave a few claps (it means a lot) and leave a follow if you enjoy my other content too.
Are you interested in learning about what goes behind Deepfakes or how Quantum Computing actually works? Then feel free to check out my other articles!
See you in the next one!
