Slides 2020

• Week 1-a: course organisation, what’s NLP about, overview of problems, first contact with text data
• Week 1-b: recap of probability theory, maximum likelihood estimation
  • Further reading: Notes on MLE for Bernoulli and Categorical
• Week 2a and 2b:
  • Before class: Notes on directed graphical models and slides on directed graphical models
  • Week 2-ab: probability of a sentence, n-gram language models, smoothing, perplexity
  • Textbook: Ch4 (2nd edition) or Ch3 (3rd edition)
• Week 3-a and 3-b:
  • HMMs: parts of speech, hidden Markov models (HMM) for language modelling and POS tagging, Viterbi algorithm
    • Marginal probabilities: notebook and tutorial
  • Textbook: Ch5 (2nd edition) or Ch8 (3rd edition) and Appendix A (for the forward algorithm)
• Week 4: no lectures
• Week 5
  • Feature-rich models: naive Bayes and logistic CPDs
  • Applications to classification and to generative models

Slides 2019

• Week 5
  • Probabilistic context free grammars
  • Parsing and inside algorithm: pdf, keynote
  • Textbook: Ch 12-14 (2n edition)
• Week 6-7
  • Feature-rich models: logistic CPDs
  • Applications of feature-rich models to NLP
    • text classification and richer components in generative models
• Week 8: final exam