UPCOMING EVENTS

SIG for Language & Speech Technology Development

2 nd CALL FOR PAPERS April 2007  

Special Track for Human Language Technologies (HLT)

A Special Track for Human Language Technologies (HLT) is being organised by the ALASA Special Interest Group for Language and Speech Technology Development, to run in parallel with the
14th International Conference of the African Language Association
of Southern Africa.

Dates : 9 - 11 July 2007
Venue: Nelson Mandela Metropolitan University, Port Elizabeth
Theme: HLT research in South Africa: towards sustainable development*

 *Due to practical reasons, only papers specifically focussing on languages spoken in Africa will be considered.

Sub-themes:

a) Resource development
b) Speech processing
c) Text processing
d) HLT applications

 Important Extended Dates:

30 April 2007 - Submission of abstract and summary
15 May 2007 - Notification of acceptance
Time allocated to each presentation: 20 min plus 10 min discussion

 Purpose of Special Track:
• to bring together the growing Human Language Technologies (HLT)
community in South Africa and encourage interactions;
• to encourage dissemination of results;
• to foster interaction between academic and industrial research;
• to provide a forum for discussion of new research and
students research;
• to increase visibility of HLT research in South Africa.

Submission of abstracts and summaries:
Abstracts and summaries should be submitted by e-mail as an MS Word file attachment in accordance with the following guidelines:
• Author’s title, surname and first name
• Institutional affiliation
• E-mail address
• Telephone and/or fax numbers
• Title of the paper
• An abstract of the paper of approximately 100 words
• A summary of the paper of approximately 500 – 800 words (excluding the “References”), stating the research questions and aims, the theoretical approach and the main findings of the paper.

• Please submit your abstract and summary to: boschse@unisa.ac.za

• An acknowledgement of receipt of your submission should reach you within three days. If this is not the case, please enquire as soon as possible.

 Submission of Papers:

Authors may submit their papers for publication in the South African Journal of African Languages as individual contributions. Instructions to authors are obtainable from Prof. Albert Kotze kotzeae@unisa.ac.za

Registration details and other information:

Should you require more information regarding registration and accommodation, please respond to the ALASA Conference Secretariat:

Mrs Yolande Ferreira yolande.ferreira@nmmu.ac.za

Tel: +27 41 504 2187 Fax: +27 41 5049353

More information will also be available on the website soon.

http://www.alasa.org.za/sig

See also: http://www.meraka.org.za/nhn

Special Track Organisers:

Justus Roux

Stellenbosch University Centre for Language and Speech Technology

jcr@sun.ac.za

Sonja Bosch

University of South Africa

boschse@unisa.ac.za

Danie Prinsloo

University of Pretoria

PAST EVENTS

COMPUTATIONAL MORPHOLOGY WORKSHOP

During this workshop participants will be instructed hands-on in the use of Xerox Finite-state tools and techniques to build useful and efficient programs that process text in natural languages. The emphasis will be on the indigenous languages of South Africa.

WHO SHOULD PARTICIPATE: 

• Linguists/lexicographers who intend developing computational morphological analysers for African languages

•  Computer scientists/programmers who wish to team up with linguists as described above

• Newcomers as well as those wishing to re-start or review work done during a previous Xerox Finite State workshop

The workshop forms part of the SIG Capacity Building Project which is aimed, among others at the development of morphological analysers for all the African languages of South Africa, as part of the basic language resources kit. The development of the Zulu (isiZulu); Northern Sotho (Sesotho sa Leboa) and Xhosa (isiXhosa) analysers is well under way, while development work for the Ndebele (isiNdebele) and Tswana (Setswana) analysers has already begun, following Workshops 1 and 2 of the project. Further participants or teams that have an interest in the development of analysers for languages such as Southern Sotho (Sesotho), Tsonga (Xitsonga), Venda (TshiVenda), Swati (siSwati) or other African languages are most welcome to participate.

PREREQUISITES:

Linguists/lexicographers with no programming background are advised to team up with a computer scientist/programmer, because:

• Writing a finite‑state morphological analyser is a kind of computer programming, and it is very difficult to follow the course and progress on to practical development if you have not done some kind of computer programming in the past.  It does not matter which programming language(s) you have used (C, C++, Perl, Python, Java, Prolog, Lisp, etc.), the general programming discipline and skills are re‑applicable.

•  Writing a morphological analyser also requires a detailed, conscious knowledge of orthography, morphotactics, morphophonology, lexicography, and sometimes a limited knowledge of etymology.  The people who master this knowledge are descriptive linguists, and they may not have programming experience. Conversely, the typical programmer may not have the necessary linguistic knowledge.

COURSE MANUAL

The course manual, including the applicable software, that needs to be purchased before the workshop is: 

Beesley, K.R. & Karttunen, L. (2003). Finite-State Morphology. Stanford, CA: CSLI Publications. ISBN  1-57586-434-7 

For more information on the above publication, consult:  http://www.fsmbook.com/
The publication may be ordered from:
http://www.exclusivebooks.co.za
The price quoted is: R372.00

After purchasing the book, participants should contact Lauri Karttunen at karttunen@parc.com to get the latest updates of the finite‑state code.  These updates correct a few bugs and eliminate some infelicities. The updates also handle Unicode much better.

LAPTOP

You need to install the Xerox Finite-State Software which is included as a CD in the above mentioned book, on your own laptop on which you will be working hands-on during the workshop. Instructions for installing the software are included in the book. The software is compatible with the following operating systems: Solaris, Linux, MacOS X as well as Windows.

REGISTRATION

Please e-mail your application to: boschse@unisa.ac.za
Clsoing date: 10 September 2004

REGISTRATION FEE

• R500.00 for the full four day course

•  R150.00 per day for those who have participated in a previous Xerox Finite State workshop and wish to review certain parts of the work (cf. Programme below)

Refreshments and lunch are available in the UNISA cafeteria (for your own account). 

 ACCOMMODATION

Accommodation (for your own account) is available at:

Cosmos Guest House

Bed and Breakfast:
Standard unit: R 190.00 per person single
Luxury unit: R 240 per person single

Please book your accommodation directly at: 

Tel.012-3410890
Fax 012-3412052
e-mail: guesthouses@tut.ac.za

PROGRAMME

• Monday, 20 September 2004
  Morning session:   
  Gentle introduction to finite-state machines: concepts, terminology
  Afternoon session: 
  Introduction to the xfst interface language and regular expressions

• Tuesday, 21 September 2004
  Morning session:
  More on xfst: defining lexical transducers 
  Afternoon session: 
  Introduction to the lexc language (used for defining dictionaries and morphotactics)

• Wednesday, 22 September 2004
  Morning session:
  A more formal introduction to finite-state languages, relations, and machines
  Afternoon session: 
  Developing the finite-state mindset: composition is our friend

• Thursday, 23 September 2004
  Morning session:
  How to test systems using the finite-state languages
  Afternoon session: 
  How to use finite-state filtering and flag diacritics

CONTACT

For further information please contact:

Prof. Sonja Bosch: Project Coordinator
Department of African Languages
PO Box 392
0003 UNISA

Tel. 012 429 8253
Fax 012 429 3355
Cell 082 55 88 586
e-mail: boschse@unisa.ac.za

WORKSHOP ON STANDARDS FOR LANGUAGE RESOURCES
16 & 17 March 2004
University of South Africa, Pretoria

PROGRAMME
Session 1: Introduction to the ISO context on language resource management
Overview of ISO TC 37/SC 4, current activities, background on the work achieved with ISO 16642 (TMF)
Discussion: Applicability of the activities of ISO TC 37/SC 4 to the South African context 

Session 2: Morpho-syntactic annotation
General principles of morpho-syntactic tagging. Current work: the MAF
(Morpho-syntactic Annotation Framework, ISO WD 24611) model.
Discussion/work: Short presentations on existing practices regarding morpho-syntactic annotation in South African languages (existing tagsets, existing taggers, existing documentation). 

Session 3: Data categories
Introduction to the main ideas behind the deployment of a Data
Category Registry for ISO TC 37
Presentation of the on-line tools for managing data categories
Hands-on session on describing data categories
Discussion: Possible contributions to the DCR from a South African perspective 

Session 4: Lexica
Current status of activities within SC 4 with the LMF (Lexical Markup
Framework), as discussed in early February 2004: global model,
application to the description of a morphological lexicon
Working session: Short presentations by participants on their own lexical data, their problems etc.
Conclusion: The way forward – Justus Roux (Chairman, STANSA TC 37) 

PRESENTERS

Laurent Romary
Laurent Romary obtained his PhD in computational linguistics in 1989 and completed his Habilitation thesis in 1999. He is currently senior researcher at INRIA (Institut National de Recherche en Informatique et Automatique), France. He leads the Langue et Dialogue team (http://www.loria.fr/equipes/led/) at
Loria laboratory and has conducted various research activities on man-machine dialogue, multilingual document management and linguistic engineering. His has participated in numerous national and international projects related to the representation and dissemination of language resources and on man-machine interaction, and in particular coordinated the MLIS/DHYDRO and IST/MIAMM projects. He has been the editor of ISO 16642 (TMF – Terminological Markup Framework) under TC37/SC3, and is the Chairman of ISO Committee (TC37/SC4) on Language Resource Management.

Philippe Sebire
Philippe Sebire, who is also affiliated to INRIA, is an engineer supervising the development of the tools that are being put together for the management of data categories in ISO TC 37. He manages a team which develops tools for management of a data category registry (in relation with ISO12620 and ISO16642) and for submitting new data categories in a ISO process. Philippe Sebire's visit is sponsored by INRIA.

===========================

ALASA-SIG & AST 2003 WORKSHOP
Human Language Technologies: Research and Practice
7- 8 July 2003
Stellenbosch, South Africa

PROGRAMME

OUTLINE AND SCOPE OF HLTS
Justus Roux, African Speech Technology Project
HLT Research and Development in South Africa
Etienne Barnard, University of Pretoria
Marelie Davel, CSIR
Human language technologies in the developing world

LANGUAGE AND SPEECH RESOURCES
Pippa Louw, African Speech Technology Project
Annotating the AST Speech Databases: Practise makes Perfect

DEVELOPMENT OF BASIC LANGUAGE RESOURCES
Jackie Jones, Laurette Pretorius and Sonja Bosch, University of South Africa
Towards a machine-readable Xhosa lexicon using XML
Sonja Bosch and Laurette Pretorius, University of South Africa
Development and application of a Zulu XML lexicon

RESEARCH PROJECT REPORTS: COMPUTATIONAL MORPHOLOGICAL ANALYSIS
Winston Anderson, Albert Kotze & Petro du Preez, University of South Africa
A finite-state transducer for phonologically induced palatalisation: evaluating
Two approaches to labial palatalisation in Northern Sotho
Axel Fleisch, Reneé Fourie
Project “Finite State Morphological Analyser Ndebele"
Arvi Hurskainen, University of Helsinki
Louis Louwrens & George Poulos, University of South Africa
Disjoining writing as a computational problem

============================

LANGUAGE & SPEECH TECHNOLOGY DEVELOPMENT COLLOQUIUM
HLT in South Africa – development and implementation
15 October 2002
Venue: University of Stellenbosch

PROGRAMME
Justus Roux (African Speech Technology Project)
The development and implementation of Human Language Technologies (HLT) in South Africa
Ken Beesley (Xerox Research Centre Europe, Grenoble)
Finite-state morphological analysis and minority languages
Sonja Bosch (University of South Africa)
Computational morphological analysis in the context of African languages