Home > Week 2
Week 2 has 3 topics + 1 assignment. The first topic is lower priority (theory + context), but Topics 2 and 3 are the core of modern NLP - work through them carefully. Each topic builds on the last.
01 - Classical ML & Evaluation (Lower Priority)
└── Understand where the field came from, and why it wasn't enough.
02 - Introduction to Deep Learning
└── The architecture that powers everything today.
03 - Embeddings & Word2Vec
└── How machines learn to understand meaning in text.
Assignment - Sentiment Analyzer / Name Completer
└── Build something real with what you've learned.
| # | Topic | What You Will Learn | Est. Time |
|---|---|---|---|
| 1 | Classical ML & Evaluation | BoW, TF-IDF, Logistic Regression, Loss, Accuracy, F1, Confusion Matrix | 2–3 hrs |
| 2 | Introduction to Deep Learning | Neurons, Perceptrons, MLPs, Hidden Layers, ReLU, tanh, Softmax | 4–5 hrs |
| 3 | Embeddings & Word2Vec | Bigrams, One-Hot, Dense Vectors, Word2Vec (Skip-gram & CBOW) | 3–4 hrs |
- Can you explain why Bag-of-Words loses word order, and why that matters?
- Can you draw a 3-layer MLP and label every part (weights, activations, output)?
- Can you explain why one-hot encoding is a bad representation for words?
- Can you describe in your own words what Word2Vec is learning?
If yes, you are ready. See you in Week 3 →