51. What Is Natural Language Processing (NLP)
51.2 Classical NLP vs ChatGPT Lanes
Why. Beginners mix three words: NLP, ML on text, and ChatGPT. Clear labels stop wrong tool choice.
What. NLP is the field. Classical NLP = rules + bag-of-words / TF-IDF + sklearn classifiers. Deep NLP = embeddings / transformers (you saw a tiny taste in ch49–ch50). Generative AI = models that write new text (LLM APIs) — covered mainly in the Gen AI course.
How.
Pick the right lane
- Hundreds of labelled Mauli reviews, need positive/negative → classical NLP (ch52–ch53).
- Need a chatbot that drafts WhatsApp replies → Gen AI / LLM.
- Need embeddings for semantic search at scale → Gen AI embeddings + vector DB (not this chapter).
lanes = {
"classical_nlp": "TF-IDF + LogisticRegression on reviews",
"tiny_dl_text": "Embedding mean-pool (ch50 style)",
"gen_ai": "LLM API for draft replies / RAG",
}
for k, v in lanes.items():
print(f"{k}: {v}")
Course map
This Data Science block teaches classical NLP. Gen AI guides already cover LLMs, RAG, and embeddings on the live Gen AI course.