Tech It Out: Natural Language Processing

Image of a computer screen with code

Let’s zoom in to a very crucial part of human intelligence: Language. The study of language is an extremely interesting field because it involves so many aspects of human intelligence and communication. Understanding and generating language involves memory retrieval, speech recognition, planning and a great deal of reasoning, among many other functions. For example, we often plan our sentence construction (even when this happens extremely fast), we reason about world events and use those facts strategically in conversation, we are able to understand the language when spoken in different accents and we can understand when someone is using language ironically (some more than others). But, computers are not actually able to understand human or Natural Language. In fact, computers operate by storing numerical data and performing calculations. So how does Siri, Alexa or Cortana understand you? This is where the field of Natural Language Processing (NLP) comes in.


What is NLP?

Natural Language Processing is essentially the field that deals with how to process and analyze human language so that computers can understand, process and produce it. Of course, there isn’t a single “Language Module” which can perform all language tasks the way humans can. Just as the chess example shows a very specific planning function, NLP models can often perform one or a handful of tasks. For example, a machine translation system such as Google Translate is able to translate from one language to another but will not be able to answer you if you ask it “where is the nearest supermarket?’” You would need a question answering or Information Retrieval system. Although these are all NLP systems, they are trained to ultimately perform different tasks. 


NLP is a part of many people’s lives, which is precisely what drove me to be interested in the field. It may not be something people think about every day, but behind technologies such as Google Translate, personal assistants (such as Siri, Cortana and Alexa), grammar checking tools in any text editor and your favorite search engines, there is a huge effort by many NLP researchers and engineers. Additionally, with the emergence of Machine Learning, NLP systems are becoming more accurate at classifying and generating human language. This has led many NLP researchers and practitioners to use their field beyond the applications we tend to think about.