co-ban Beginner
What is NLP (Natural Language Processing)?
The branch of AI that helps machines read, understand, and produce human language — from Google Translate to ChatGPT.
Updated: May 5, 2026 · 2 min read
NLP (Natural Language Processing) is the branch of AI focused on helping computers read, understand, generate, and interact using human language.
What can NLP do?
Understanding (NLU)
- Text classification (spam? sentiment? topic?)
- Named Entity Recognition (find names of people, places, companies)
- Question Answering
- Summarization
Generation (NLG)
- Writing articles, emails, code (LLMs)
- Machine translation
- Image captioning
- Text-to-Speech
Bridges
- Speech-to-Text (Whisper)
- OCR (reading text from images)
A short history
| Era | Characteristics | Examples |
|---|---|---|
| 1960-1990 | Rule-based | ELIZA chatbot, hand-coded translation |
| 1990-2010 | Statistical NLP | Early Google Translate |
| 2010-2017 | Word embeddings (Word2Vec, GloVe), RNN/LSTM | Translation improved significantly |
| 2017-now | Transformer | BERT, GPT, Claude |
The big bang came in 2017 with the paper “Attention Is All You Need” — the Transformer architecture changed everything.
How are NLP and LLM different?
- NLP is the entire FIELD — every way machines process language
- LLM is a TYPE OF MODEL within NLP — currently the dominant one
Modern LLMs “swallow” most traditional NLP tasks:
- Sentiment analysis? → Ask the LLM “positive or negative?”
- NER? → “List the names of people in this passage”
- Translation? → “Translate to X”
→ Many classic NLP tasks now just need a prompt to an LLM, no specialized model required.
NLP for Vietnamese
Vietnamese has some tricky characteristics:
- Word segmentation: is “học sinh” one word or two? — the machine has to know
- Tone marks: “ma/má/mà/mả/mã/mạ” all mean completely different things
- Tokenization: takes 1.5-2× more tokens than English
Vietnamese-specific resources:
- VnCoreNLP — a classic toolkit
- PhoBERT — BERT pretrained on Vietnamese
- Vinai/PhoGPT — a Vietnamese LLM
- Llama 3.x, Claude, GPT-5: support Vietnamese fairly well, though it’s not their native focus
Common NLP applications in Vietnam
- Customer-service chatbots for banks and insurance
- Social media sentiment analysis
- OCR for ID cards, contracts, invoices
- Automatic email categorization
- Internal semantic search
Related
Tags
#nlp#language