NLP for Machine Learnings
Natural language processing is teaching models to work with text: tokenisation, cleaning, bag-of-words and TF-IDF features, word and sentence embeddings, and transformer models used through the Hugging Face libraries. This unit in the NLP and Modern AI stage covers turning words into numbers and fine-tuning pretrained models for classification.
Stage: NLP and Modern AI · Part of the Machine Learning roadmap · Last reviewed 19 November 2026
What to learn in NLP
- Tokenise text and build a TF-IDF representation for a classifier
- Explain how embeddings place similar words or sentences close together
- Load a pretrained transformer and tokenizer using Hugging Face Transformers
- Fine-tune a small model for text classification on a labelled dataset
- Evaluate a text classifier with precision, recall and a confusion matrix
NLP interview questions
What is the difference between TF-IDF and word embeddings?
TF-IDF is a sparse count-based score that weights words by frequency in a document and rarity across documents, ignoring meaning. Embeddings are dense learned vectors where words or sentences with similar meaning lie close together.
What does the attention mechanism do in a transformer?
Attention lets each token compute a weighted combination of all other tokens in the sequence, with weights learned from query and key similarity. This captures long-range relationships directly and allows parallel processing, unlike step-by-step recurrent models.
Where NLP fits in the Machine Learning roadmap
NLP belongs to the NLP and Modern AI stage (beginner level), which takes about 3 hours. Language and beyond.
Other units in this stage
Back to the full Machine Learning roadmap