AME 5003 (theory) and AME 5053 (lab) — MSIS Manipal.
Lessons, slides, labs, primers, the question bank and the glossary are all there.
Labs run in Google Colab. You need a Google account and a browser — nothing to install.
| Lab | Week | ||
|---|---|---|---|
| 1 | 1 | Refresher and setup |  |
| 2 | 2 | Information extraction with regular expressions | |
| 3 | 3 | Tokenize, normalize, stem, remove stop words | |
| 4 | 4 | Incidence matrix and inverted index | |
| 5 | 5 | N-grams and n-gram probabilities | |
| 5 | 5 | N-grams and n-gram probabilities | |
| 6 | 6 | Sentiment analysis with Naïve Bayes | |
| 7 | 7 | Cosine similarity and TF-IDF | |
| 8 | 8 | Word2Vec, GloVe, and the comparison | |
| 9 | 9 | Sequence models: RNN, LSTM and GRU on Shakespeare | |
| 10 | 10 | Attention, and a translation system | |
| 11 | 11 | Encoder and decoder transformers | |
| 12 | 12 | Final revision | |
Two things that catch everyone out:
- Save your own copy. The moment a lab opens, do File → Save a copy in Drive. Work in that copy. If you type into the shared notebook, your work disappears when the tab closes.
- Run the install cell first, every time. Colab wipes the machine between sessions, so nothing you installed last week is still there. This is normal, not a bug.
Short notebooks on the libraries the labs use. Read one when a lab reaches for a tool you have not met; none of them is a prerequisite for the lessons.
- NLTK — primer 1
- spaCy — primer 2
- NumPy — primer 3
- NumPy, Hands On — primer 4
Each project comes three ways. The brief says what to build and why, and how it is marked. The requirements sheet is the same thing in plain English with no code, for building from an empty directory. The starter repository has the corpus loading, a command line and a test suite already written, with stubs for the parts you fill in.
Both routes are marked identically — take the starter if you would rather spend the time on the measuring and the writing, and the requirements sheet if you would rather build the whole thing yourself.
| # | project | ||
|---|---|---|---|
| 1 | A mini search engine | requirements | starter code |
classroom/demos/ holds the scripts run live in class, one per session, using the same examples
as the lesson it belongs to. They are for following along, not assessed. classroom/README.md
has the setup.
All prescribed by the syllabus. The first two are free online.
- Speech and Language Processing — Jurafsky & Martin. The main text.
- Natural Language Processing with Python — the NLTK book.
- Practical Natural Language Processing — Vajjala et al. (code)
- Natural Language Processing with PyTorch — Rao & McMahan.
- A Primer on Neural Network Models for NLP — Goldberg.
Open an issue. Genuinely — if something here is wrong or unclear, saying so helps everyone in the class.
These pages are generated. Do not send pull requests against this repo; the source lives elsewhere.