LINGUIST List 7.953

Sat Jun 29 1996

FYI: MSc at Sheffield in Language and Speech Processing

Editor for this issue: T. Daniel Seely <dseelyemunix.emich.edu>


Directory

  1. Yorick Wilks, MSc at Sheffield in Language and Speech Processing

Message 1: MSc at Sheffield in Language and Speech Processing

Date: Wed, 26 Jun 1996 12:03:08 -0000
From: Yorick Wilks <yorickdcs.shef.ac.uk>
Subject: MSc at Sheffield in Language and Speech Processing

 University of Sheffield
 Department of Computer Science

 MASTER OF SCIENCE IN LANGUAGE, SPEECH AND AUDITORY PROCESSING



Introduction

This advanced MSc programme provides a sound professional education and
research training in new areas of information technology concerned with
computer perception and processing of human language in all its forms. It is
designed to provide an academic and practical grounding in part of what is
known in Europe as the Language Industries. It aims to provide training for
further research in this rapidly growing field in this Department or
elsewhere.

Language, speech and auditory processing is an inherently interdisciplinary
field, involving elements of linguistics, phonetics, computer science, signal
processing and artificial intelligence. Graduates generally come into the
field with training in a subset of these disciplines, which will vary from
person to person. One role of this Master's degree is to fill out the profile
of each student in the areas which are appropriate for that person. We
therefore aim for a wide choice of modules which can be tailored to
individual needs.

The course also provides skills in demand in today's world of language and
information in electronic publishing, political/economic and scientific
information handling, computer aids to translation, speech technology,
composition, language learning, and legal retrieval and information handling
etc.


The Academic Profile

The Department has a substantial research base in these areas, which has now
resulted in University funding for ILASH: the Institute for Language Speech
and Hearing, with which the MSc is associated. ILASH has its own computing
facilities, premises and support staff, and academic staff attached to it from
nine departments. Sheffield is a node on the EU-funded ELSNET (European
Network in Language and Speech) network and participates in many Europe-wide
programmes that give opportunities to link to work across the Community. We
are coordinating the 10-laboratory Human Capital and Mobility (HCM) EU network
SPHERE: `Representations in Speech and Hearing'. We also participate in
EU-ERASMUS programmes in speech and language where students can complete their
dissertations abroad.


Staff

The course teaching will draw on staff in Computer Science and other
departments in the University. The following is a list of current Computer
Science academic staff working in Language, Speech and Hearing together with
their research interests:

GUY BROWN:
	Auditory models, sound source separation, audition, speech perception
MARTIN COOKE:
	Auditory models, sound source separation, audition, speech perception
ROBERT GAIZAUSKAS:
	Logical models of natural language texts, information extraction from
 	corpora
PHIL GREEN:
	Speech perception, automatic speech recognition.
MARK HEPPLE:
	Computational linguistics, grammatical formalisms, parsing,
	categorial grammar
MIKE HOLCOMBE:
	Formal models of NLP, formal models of user modelling, visual formal
	specification languages
JIM McGREGOR:
	User modelling, parsing, Prolog, tutoring systems
PAUL Mc KEVITT:
	Pragmatics, intentions, natural language dialogue, revision in
	dialogue, user-computer interfaces, hyper/multimedia, user modelling,
	integration of speech, language and vision processing
STEVE RENALS:
	Large-vocabulary speech recognition, language modelling, neural nets
AMANDA SHARKEY:
	Connectionist and cognitive models of language, language acquisition,
	symbol grounding, parsing, translation.
NOEL SHARKEY:
	Connectionist Natural Language Processing, Neural Network models of
	Cognition, Neural Representations underlying language and thought, Sens
ory and
	Action grounding of concepts.
TONY SIMONS:
	Machine translation, syntactic, chart, and object-oriented parsing
YORICK WILKS:
	Artificial intelligence, natural language understanding, belief
	pragmatics, lexical computation, parsing, information extraction from c
orpora.


Entrance Requirements

Applicants will normally be expected to have, or be expected to obtain
before joining the programme, a first degree with a reasonable grade (UK 2-2
or better) in any subject, but those with degrees in computing, mathematics,
psychology, physics, electrical engineering, linguistics, phonetics or
cognitive science will be preferred. Work in an information service, computer
department, advanced publishing environment or anything similar is considered
 advantageous, but candidates without such experience will be given equal
consideration. International student applicants whose first language is not
English will be required to provide evidence of English language competence.


Structure and Content

The course consists of a taught part for two University Semesters, followed
by written examinations and then a project examined by dissertation and oral
examination. The taught part of the course will consist of twelve modules. (A
module occupies 1 semester and typically breaks down into 20 lecture hours and
10 practical/tutorial hours). Since we aim to cater for students coming from
multidisciplinary backgrounds, we endeavour to make the course as flexible as
possible. Students take six core modules and choose six electives. The advice
and approval of tutors must be sought before deciding on the choice of
elective. The six core modules are

COM333 and COM334: Natural Language Processing
COM325 and COM326: Speech and Hearing
COM411 and COM412: Research topics in Speech, and Language and Hearing

The latter consists of a series of guest lectures and local seminars which
students must attend, discuss, analyse and write essays about. Such studies
are valuable both for their technical content and for the development of
research skills, since understanding the research of others is a valuable
asset which requires practice. Students taking those modules will be expected
to take part in one of six or more research groups so as to experience
research methods and work in progress. More detail of these modules is given
below.

The elective modules offered from year to year depend upon the availability
of staff and the trends in research and professional practice. Some elective
modules are shown in the table.

The period from June to 31st August will be devoted to the preparation of a
supervised dissertation to be submitted on or before the end of the academic
year.





Core course descriptions

COM325 Computer Speech and Hearing
(Half Module - Autumn Semester)

This half-module aims to demonstrate why computer speech processing is an
important and difficult problem, to investigate the representation of speech
in the articulatory, acoustic and auditory domains, and to illustrate
computational approaches to speech parameter extraction. It examines both the
production and perception of speech, taking a multi-disciplinary approach
(drawing on linguistics, phonetics, psycho-acoustics, etc.). It introduces
sufficient digital signal processing (linear systems theory, Fourier
transforms) to motivate speech parameter extraction techniques (e.g. pitch and
formant tracking). The half-module covers state-of-the-art computational
models of auditory processing and their application to sound source
separation.

COM326 Speech Technology
(Half Module - Spring Semester)

This half-module aims to study the principles of the emergent field of speech
technology and to study typical applications of these principles and assess
the state of the art in this area. Students will learn the prevailing
techniques of automatic speech recognition (dynamic time warping, hidden
Markov models and neural networks); will see how speech synthesis and
text-to-speech methods are deployed in spoken language systems, and will
discuss the current limitations of such devices. The half-module will include
project work involving the implementation and assessment of a speech
technology device.

COM333 Natural Language Processing I
(Half Module - Autumn Semester)

This half-module aims to give a well-rounded introduction to the field of
computer processing of natural language, by exploring computational approaches
to syntax (structure), semantics (meaning), and pragmatics (context). We will
cover the standard theories, models and algorithms, discussing competing
solutions to problems, describing example systems and applications, and
introducing new and interesting research areas.
COM334 Natural Language Processing II
(Half Module - Spring Semester)

This half-module explores some of the recent developments in natural language
processing by computer, and in particular the new wave of statistical and
other empirical methods now being applied that make language processing
systems larger and more practical. The development of this line of work is
traced through machine translation, information extraction from texts, and
extraction of semantic information from dictionaries up to recent developments
in pragmatics. Techniques such as automatic part-of-speech and word-sense
tagging of texts are discussed in detail, as well as the development of
grammars empirically from texts.

COM411 Research Topics in Speech, Language and Hearing 1
(Half Module - Autumn Semester)

This half-module will be based around a series of guest lectures, local
seminars, videos etc. which students will attend, discuss, analyse and write
essays about. In each semester the student will participate in at least one
"themed" study group with researchers and academic staff: reading and
commenting on state-of-the-art papers, producing analyses and syntheses of
extended background material etc. There will also be a weekly tutorial
back-up. Such studies are valuable both for their technical content and for
the development of research skills, since understanding the research of others
is a valuable asset which requires practice, and this method will induct a
student effectively into group research methodology in computational language,
speech and hearing.

COM412 Research Topics in Speech, Language and Hearing II
(Half Module - Spring Semester)

This half-module follows on from COM411. It too will be based around a series
of guest lectures, local seminars, videos etc. which students will attend,
discuss, analyse and write essays about.

Assessment

Students will be required to pass continuous assessment and examinations for
all twelve modules, and produce an acceptable dissertation. These three
hurdles will be independent, in that to pass a student must pass all of them
and to get a distinction a student must at least approach distinction standard
in all of the continuous assessment, the examinations and the dissertation.

WHILST EVERY EFFORT IS MADE TO ENSURE THE ACCURACY OF THIS PUBLICATION, THE
UNIVERSITY CAN ACCEPT NO RESPONSIBILITY FOR ERRORS OR OMISSIONS.

For further information and an application form, please contact:

The Postgraduate Admissions Officer
Department of Computer Science
University of Sheffield
Regent Court
211 Portobello Street
Sheffield S1 4DP
UNITED KINGDOM

Phone: (+44) 0114 2825590
Fax: (+44) 0114 2780972
Email: deptdcs.shef.ac.uk
www: http://www.dcs.shef.ac.uk/teaching/msc/
Mail to author|Respond to list|Read more issues|LINGUIST home page|Top of issue