Skip to content
LiveNext 02:05:22

Definition

What is Natural language processing?

Natural language processing is Natural language processing, or NLP, is the branch of AI that lets computers read, interpret and generate human language, covering tasks such as translation, summarizing and question answering.

What NLP is

Natural language processing, or NLP, is the field of AI focused on human language. Its goal is to let software work with text and speech the way people do: understanding what was said, finding what matters and responding in kind.

Language is hard for computers. Words have several meanings, sentences depend on context, and people leave much unsaid. "Book a table" and "read the book" use the same word for different things. NLP is the work of handling that ambiguity.

Typical NLP tasks

  • Translation: converting text between languages.
  • Summarization: shortening a document while keeping the point.
  • Text classification: sorting messages into categories, such as spam or support request.
  • Speech recognition: turning audio into text, and text-to-speech for the reverse.
  • Question answering and search: finding or composing an answer from text.
  • Generation: writing new text, which is what a chatbot does.

How the approach has changed

Early NLP used hand-written grammar rules and word lists. Later, statistical machine learning let systems learn patterns from many examples. Then deep learning took over, and the transformer design, built on an attention mechanism, made it possible to train very large models on huge amounts of text.

Text is first broken into pieces by a tokenizer, and each piece is turned into an embedding, a list of numbers that captures meaning so similar ideas sit close together.

Why it matters to you

Nearly every language feature you touch uses NLP: autocomplete, voice assistants, email filters, translation apps and AI search. Today a single large language model can perform many of these tasks that once needed separate specialized systems, which is why the line between NLP and general AI has blurred.

An example

A support team receives thousands of emails a day. An NLP system reads each one, labels it as billing, bug or cancellation, pulls out the product name and drafts a reply for a person to approve. No one wrote rules for each phrasing. The model learned from past tickets.

How it relates to neighboring terms

NLP is a field and a large language model is one tool within it. Computer vision handles images in the same way NLP handles words, and a multimodal model handles both.

Questions people ask

What is the difference between NLP and an LLM?

NLP is the whole field of getting computers to work with human language. A large language model is one powerful technique within that field.