Breaking Down NLP- Neuro-linguistic Programming

There were times when you had to type everything you want a computer to do but AI engineers decided to turn things easier. The convenience of commanding Siri or Alexa to do your tasks for you and transmitting your instructions to the servers is just one of the many wonders of AI, and primarily, of NLP.

NLP (Neuro-linguistic programming) is a merge of different fields like linguistics, computer science, information engineering, and artificial intelligence. The purpose of the development of NLP was to make it possible for the computers to read, understand, and perceive the Natural languages of humans and communicate with them. Here’s a brief overview of how this technology works.

The process of NLP is not as easy as it sounds. AI and ML engineers have to solve this problem in several steps and then chain them together to make it all work.

The first step is to fragment the text into separate sentences from a paragraph and then process one sentence at a time. To make it more detailed the sentences are then fragmented into separate words often referred to as tokens by engineers. Next, the computer tries to understand the basic form of each word by a process called lemmatization. After this goes to filter out the stop words (words that you might want to filter out before doing any statistical analysis.)

After the computer has fully understood every word it works to integrate them by understanding the relations between the words. Now that the computer knows each word as a separate entity and their relations as well it tries to make sense and extract ideas out of it by Named Entity Recognition (NER). Since the computer figured each sentence separately, to make meaning of the whole paragraph and the pronouns used in it has to do Coreference Resolution which involves mapping by tracking pronouns across sentences and find out the words that refer to the same object.

With all the above NLP steps done Siri is finally able to process your google search instructions.