NLP

Natural Language Processing (NLP) is the name for the branch of AI that deals with "the language people use." Human language is hard for machines because its rules are loose and meaning shifts with context — we naturally know whether "bat" means an animal or a piece of sports equipment, but a computer has to learn that.

Translators, spell checkers, spam filters, voice assistants — any service that understands or handles language runs on NLP technology. Large language models (LLMs) like ChatGPT and Claude can be seen as NLP's newest and most general-purpose achievement.

When news articles mention "the NLP field" or "natural language understanding," read it as referring to the broad set of technologies that let AI handle human language. It's one of the two major branches of AI, standing alongside computer vision, which handles images.

"NLP-based sentiment analysis automatically categorizes customer complaints" — this means the machine reads the nuance in sentences written by people.

1. Try typing this into a chatbot: "Classify the sentiment of these three sentences as positive, negative, or neutral: (1) This cafe has the best vibe (2) The delivery was way too late (3) It's cloudy today."
2. The machine reads the context and tone to classify them — this is "sentiment analysis," one of the most basic tasks in NLP.
3. One step further: ask the same chatbot to "translate the sentences above into Japanese and French." Translation is another classic NLP task.