InternshipKey terms and common mistakes
The words you will hear in Natural Language Processing work and the mistakes that most often cost interns approvals.
Key terms
| Term | Meaning |
|---|---|
| NLP | Natural Language Processing |
| Tokenisation | Splitting text into tokens |
| Bag of Words | Word count representation |
| Spam filter | Classifies emails as spam or not |
| NER | Named Entity Recognition |
| Seq2seq | Sequence-to-sequence model |
| BERT | Bidirectional encoder model |
| LLM | Large Language Model |
Common mistakes to avoid
- Testing on the same text used for training.
- Cleaning away words that carry meaning, like "not".
- Reporting accuracy on unbalanced classes instead of F1.
- Ignoring mixed-language text like Tanglish.
- Keeping names and phone numbers in training data.
Want to go deeper? The free Natural Language Processing learning path has short modules, games and a certificate for each module.