NLP is a subset of AI and uses ML/DL techniques.
In computer science, languages that humans use to communicate are called "natural languages". Examples include English, French, and Spanish. Early computers were designed to solve equations and process numbers. They were not meant to understand natural languages.
NLP is rooted in the theory of linguistics. Techniques from machine learning and deep neural networks have also been successfully applied to NLP problems. While many practical applications of NLP already exist, NLP has many unsolved problems.
Sentiment Analysis From product reviews or social media messages, the task is to figure out if the sentiment is positive, neutral or negative. This is useful for customer support, engineering and marketing departments.
Machine Translation: Suppose original content is published only in one language, machine translation can deliver this content to a wider readership. Tourists can use machine translation to communicate in a foreign country.
Question Answering: Given a question, an NLP engine leveraging a vast body of knowledge, can provide answers. This can help researchers and journalists. Whitepapers and reports can be written faster.
Text Summarization: NLP can be tasked to summarize a long essay or an entire book. It can provide a balanced summary of a story published at different websites with different points of view.
Text Classification: NLP can classify news stories by domain or detect email spam.
Text-to-Speech: This is an essential aspect of voice assistants. Audiobooks can be created for the visually impaired. Public announcements can be made.
Speech Recognition: Opposite of text-to-speech, this creates a textual representation of speech.
NLP is broadly made of two parts:
Natural Language Understanding (NLU): This involves converting speech or text into useful representations on which analysis can be performed. The goal is to resolve ambiguities, obtain context and understand the meaning of what's being said. Some say NLP is about text parsing and syntactic processing while NLU is about semantic relationships and meaning. NLU tackles the complexities of language beyond the basic sentence structure.
Natural Language Generation (NLG): Given an internal representation, this involves selecting the right words, forming phrases and sentences. Sentences need to ordered so that information is conveyed correctly.