Natural Language Processing
NLP is AI for human language: text or speech. Levels: tokens → syntax → meaning → dialogue. Tasks: classification (spam), translation, NER, Q&A, summarisation, chat. Today: transformers / large language models. Still fail on sarcasm, rare words, and invented facts (hallucination).
Tiny start: split “Asha got 90 marks” into tokens. That is not understanding yet. Don’t jump to GPT without saying tokens and a task.
Natural Language Processing — output — ['asha', 'got', '90', 'marks'] then a reminder. Classification/chat comes after.
text
│ tokenize
▼
tokens → task
(class / translate / chat)One task + one failure mode.