Natural Language Processing — Part 1
NLP has been growing at a reasonable pace alongside the advances in mathematics and Deep Learning. However, there’s more to it than just text based pre-processing and simple computations. With this article, let’s dive into what NLP is and what it isn’t. Buckle Up!
Going by the textbook definition, NLP is a field of Computer Science , using up the brains from AI and Deep learning mechanisms to bridge the gap between systems and humans, when it comes to natural languages that we converse in.
Now comes the main question, why is this field not so popular yet? Well, human language is a complex composition of multiple language modifications, imports from other languages and some slang that we use in day to day conversations. As humans, we rely heavily on context and base our replies based on the context and environment that we’re subject to. However, a system with Random Access Memory has no information or storage from last computed states or operations conducted. So if you asked it to suggest or complete a sentence, it’s definitely not going to give the right output.
Moreover, as beings of emotion, we use a lot of statements in our conversations which are very context reliant. For example, idioms or metaphors, as humans we definitely get the meaning and implication. A system wouldn’t. To them, it’s yet another statement , a simple sentence.
A few of the NLP applications are :
- Massive management of textual information sources for :
a. Human use.
b. Automatic collection of linguistic resources.
2. Person/Machine interaction.
3. Massive management of textual information resources such as :
a. Machine Translation.
b. Information Retrieval.
c. Question Answering.
d. Information Extraction and Summarization.
We shall try covering these in depth in the coming articles. For now we shall leave it here for the applications.
You must be wondering, how did this field come to exist? I mean, with the influx of huge amounts of data, data driven fields such as Data science and Cloud computing, along with predictive methodologies and statistics gave rise to Artificial Intelligence and a whole new plethora of stuff like Machine learning and Deep learning. How did NLP spring up? Well, the history traces back to the year 1950 when Alan Turing published an article on Machine and Intelligence. A couple of years later, there were numerous researches going on about converting Russian into English. One of the core subjects of any Computer Science graduate would be Theory of Computation. This subject explored the general language rules, grammatical rules and Finite State automata.
NLP involves the use of FSA and transducers to map out various forms of a word. For example, the word talk can be the root word for talking, talks, talked. For a small instance like this, one would have 3 FSAs to map out the forms of the root word. To check the grammatical correctness of a sentence, the use of a syntax tree is also employed. Chomsky and his team’s work was quite instrumental in strengthening NLP’s fundamentals. More on that in a later article.
So that ends our first introductory article on NLP. Feel free to drop down comments to reach out to us. See you in the next release. You could also let us know what you’re looking forward to seeing in the next release.
Team FACE.
