Skip to navigation | Skip to main content | Skip to footer

COMP37412: Dialogue Systems (2007-2008)

This is an archived syllabus from 2007-2008

Dialogue Systems
Level: 3
Credit rating: 10
Pre-requisites: No Pre-requisites
Co-requisites: No Co-requisites
Lecturers: Allan Ramsay
Course lecturer: Allan Ramsay

Additional staff: view all staff
Sem 2 w19-25,29-32 Lecture 1.1 Fri 14:00 - 16:00 -
Assessment Breakdown
Exam: 80%
Coursework: 20%
Lab: 0%
Degrees for which this unit is optional
  • Artificial Intelligence BSc (Hons)


The aim of the course is to provide students with the knowledge and practical skills required for developing natural language interfaces to a variety of back-end applications. Dialogue systems require a mixture of technologies -- speech recognition and generation, natural language understanding, and inference about the user's beliefs and goals. The course will outline the architecture of a typical dialogue system, explaining the roles of the various components; describe the theoretical basis of each component; and consider the limitations of existing techniques. Students will have an opportunity to experiment with state-of-the-art technology, in order to familiarise themselves with the major techniques and with the problems involved in developing practical systems.

This course is complementary to COMP30421: Natural Language Engineering, and has a small degree of overlap with COMP20442: Artificial Intelligence Programming. Neither of these course units is a pre- or co-requisite, and this unit is designed to be taken by students who have not taken either of the others. Student who have taken one (or both) of these will therefore find that some material is repeated.

Learning Outcomes

On successful completion of this course unit, students should be able to:

Academic knowledge

Describe the architecture of a typical dialogue system, and explain the roles of the various components
Demonstrate an understanding of the principles underlying each component
Explain the limitations of such systems, and discuss how to evaluate the effectiveness of individual components and complete systems

Subject practical skills

Develop simple speech driven dialogue systems to provide access to back-end information systems

Transferrable skills

Understand the issues involved in assessing software systems with 'soft' requirements

Assessment of Learning outcomes

Learning outcomes (1)-(3) are assessed by examination. Learning outcomes (4) and (5) are asssessed by coursework.


Introduction: applications, architectures, limitations. (1 lecture)

Speech recognition & generation: characteristics of speech; the vocal tract; representing speech signals; hidden Markov models; formant vs diphone based synthesis. (2 lectures)

Grammar: introduction to context-free grammars; general purpose grammars vs. dialogue grammars; using Wizard-of-Oz techniques for obtaining dialogue grammars. (2 lectures)

Meanings: grammar annotation; slot & filler representations; full logical forms, compositionality (2 lectures)

Discourse structure: dialogue grammars, coherence, text generation (2 lectures)

Mixed initiative dialogues: planning, epistemic reasoning (2 lectures)

Integration & evaluation. (1 lecture)

Practicals: (1) analysis of speech signals, (2) building a dialogue system

Reading List

Core Text
Title: Speech and language processing: an introduction to natural language processing, computational linguistics, and speech recognition (2nd edition)
Author: Jurafsky, Daniel and James H. Martin
ISBN: 9780135041963
Publisher: Pearson International
Year: 2009