Pensum & examples of exam questions

Here is a list of sections from Martin & Jurafsky (second edition) that covers the dialogue systems part of the course:

- section 2.2.1: basic definition of a finite-state automata (used for dialogue management)

- section 3.11: minimum edit distance (used for e.g. speech recognition evaluation)

- section 4.2: N-gram models (used for language modelling in speech recognition)

- sections 6.1 and 6.2: basic probabilistic modelling with Markov Chains and Hidden Markov Models

- section 7.1: phonetic transcription

- section 7.2 and 7.3: basics of articulatory phonetics and pronunciation variation (the general principles, not all the details)

- section 7.4: basics of acoustic phonetics (idem)

- sections 8.1, 8.2, 8.3 and 8.4: read once, but without going into all the details

- section 8.5: understand how unit selection works and how the constraints are realised

- section 9.1 and 9.2: important concepts to understand!

- section 9.3 and 9.4: basic principles, but not the details

- section 9.8: how WER is calculated

- section 12.8: how spoken language syntax differs from written language

- section 17.6: embodiment and situation (links to our discussion of situated language processing)

- sections 21.3, 21.4 and 21.6.3: how reference resolution works

- Finally, the whole Chapter 24 should be read.

You can also find here a list of example questions that was prepared for the exam in 2012.  Note that the course given 2 years ago had a stronger focus on probabilistic modelling than this year.  Question 7 may be therefore a bit difficult for you to answer, but it is in any case a good exercise!