Translating and the Computer 30 Conference

Day One: 27 November 2008
Go to Day Two

08.30 Registration

09.15 Introduction by Chair: Chris Pyne, SAP AG, Germany

09.20 TikiWiki: New Exciting Multilingual Capabilities
Marc Laporte, AvanTech.net, Canada

10.00 WeBiText: Building Large Heterogeneous Translation Memories from Parallel Web Content – Alain Désilets, National Research Council of Canada
This paper reports on experiments carried out to evaluate the feasibility and usefulness of   a generic, universal, very large, heterogeneous Translation Memory based on parallel text mined from the web.  This concept has been named WeBiText.  A unique feature of   WeBiText is that it does not rely on corpus donations and uses data that is already posted on the web.  The advantage is that it enables the development of a very large TM without    having to first get a critical mass of organizations to contribute corpuses.  The downside is that the data used by WeBiText is much noisier, since it has not been carefully massaged      for inclusion into a TM.  The next stage in the product's development is to carry out a test with pilot users.  It is expected that this will uncover the needs for new features which were not necessary in traditional TMs, but become invaluable with a large TM developed form uncontrolled heterogeneous sources.  Another area of research relates to targeted crawling of sites with parallel content.

10.40 Discussion

10.50 Coffee

11.15 Hybrid Architectures for Multi-Engine Machine Translation – Andreas Eisele, Saarland University, Germany
This paper describes different architectures that combine rule-based and statistical   machine translation (RBMT and SMT) engines into hybrid systems. One of them allows combining many existing MT engines in a multi-engine setup, which can be done under the control of a decoder for SMT. Another architecture, uses lexical entries induced via SMT           technology to be included in a rule-based system. For all these approaches prototypical    implementations have been done within the EuroMatrix project and some indicative      results from the recent evaluation campaign are given, which help to highlight the    strengths and weaknesses of these approaches.

11.50 Discussion

11.55 CLIR-Based Collaborative Construction of Multilingual Terminological Dictionary for Cultural Resources – Mohammad Daoud, National Institute of Informatics, Japan and Joseph Fourier University, France
This paper will describe the construction of a CLIR-based multilingual       terminological dictionary dedicated to the Silk Road.  Since the beginning of the Digital Silk Road Project conducted at NII a huge amount of cultural resources has   been digitized, including 95 rare books written in 10 different languages.  As these books are important references it is desirable to make them searchable, with across lingual information retrieval system.  As an experiment, 100 000 monolingual key terms were translated from the digitized historical books.  After suitable filtration, each term was extracted into 12 languages automatically from online multilingual encyclopedias, glossaries and other online resources.  In the     next phase of development, new languages will be added and particularly attention      will be given to languages along the Silk Road, especially less resourced languages and "frozen" languages like Sanskrit.

12.30 Discussion

12.35    The African Network for Localisation (ANLOC) Manal Ahmin, ANLOC, Morocco
This project addresses the issues of localisation and local languages in ICTs through a number of sub-projects. Each sub-project is chosen to address a specific localisation need whether enabling (a    once off task designed to allow digital work in a language) or localisation (the actual work of        making a piece of software work in a local language). The overall objective of the research Project is to: a) Build an African network of localizers; b) Lower and eliminate participatory barriers: make sure that technology limitations for most African languages are eliminated; c) Increase technology and skills to facilitate localization: build           technology and increase skills within Africa that facilitates   increasing localization effort; d) Enhance allied localization technology: build tools and increase resources for technologies that assist the process of localization; and e) Influence policy: examine ICT and language policy within Africa and use outcomes from the other sub-projects to inform and influence policy makers. These objectives will be achieved through 10 subprojects as follows: Locales Development, Keyboards Development, Fonts Development, Spell Checkers Development, Terminology Development, Localization Tools Development, Training on Localisation, Localisation Software, Language and ICT Policy, and Networking     Activities.

13.10  Discussion

13.15  Lunch and Exhibition

14.30  Introduction by Chair: Reinhard Schäler, University of Limerick, Ireland

14.30  Panel Discussion – the last 30 years - Moderator: Olaf-Michael Stefanov, Vienna
A Look Back: Breakthroughs of the past 30 years - the ones that were, the ones that weren't - and why; lessons learned, and a few surprises

15.30  Tea break

15.45   Zen and the Art of Quality Assurance: Quality Assurance Automation in Translation: Needs, Reality and Expectations
Julia Makoushina, Palex Languages and Software, Russia and Hendrik J. Kockaert, Lessius University College, Belgium
This presentation follows on from the presentation given at last year’s conference by Julia Makoushina on translation quality assurance practices and tools.  While the previous presentation mainly highlighted the acceptance and usage of existing automation tools in the translation community, this presentation is more comprehensive.  It embraces all stages of the translation process and quality assurance needs of each of them. It will consider the tasks that are necessary to ensure quality as each stage and identify the ways to automate them.  At the end of the presentation, a prototype software suite will be presented, a product which is expected to overcome most of the limitations in computer assisted TQA tools   that have been surveyed to date.

16.15 Discussion

16.20 Tools, Tools, Tools – The lack of interoperability, standards and shared resources in an automated translation process

- Operability does exist: a view – tbc
- Operability does not exist: a view – Sven C. Andrä, Andrä AG, Germany

16.45  Discussion

17.00  End of Day One. You are invited to:
          30th Conference Drinks Reception

17.30-19.30
LTC invites you to attend:                        

Computer-assisted Project Management  - an overview  
LTC are pleased to invite you to join us as we show you how you can use   technology for computer-assisted project management. With LTC Worx, our cutting-edge multilingual process management system, we’ll demonstrate how you can boost your productivity and reduce administrative costs. A range of tools help you to manage projects efficiently in keeping on top of your finances and planning ahead effectively. We’ll show you how LTC Worx helps you centralize and manage all your crucial business data, and enables you to easily manage your resources. With LTC Worx, project management has never been so simple.

---

08.45  Coffee and Registration (for Day Two delegates only)

09.15  Introduction by Chair: Olaf-Michael Stefanov, Vienna

09.20 Ulla Falk-Petersen, Director, European Court of Auditors

10.00 Discussion     

10.05  Wiki behind the Firewall – Micro Scale Online Collaboration in a Translation Agency – Dave Calvert, TransForm, Germany
This presentation provides an introduction to some of the applications that can be realized   using wiki software in a small language service provider (LSP).  It is based on experience gathered with the wiki systems in use at Transform GmbH.  A wiki is essentially a web-based, user editable database for creating, browsing through and searching for information.  It can facilitate both top-down information distribution (when a procedure is documented by management) and a bottom-up process (when a job is documented by the   employee(s) doing the work).  Both of these information flows are essential in an LSP environment.  Transform chose MediaWiki because it seemed relatively easy to install and learn.  Despite initial successes, the wiki functioned mainly as the author's personal    notepad and for top-down distribution of information.  A workshop was run to show colleagues how to create a page and this encouraged others in the company to use it. The largest weaknesses in MediaWiki are its calendar facilities and the clumsy multilingual support and so they are now testing TikiWiki.

10.40 Discussion

10.45 Coffee

11.10  Managing Complex Translation Projects Through Virtual Spaces: A Case Study  Dr Fola Yahaya, Strategic Agenda LLP, UK
Global translation projects often suffer from issues such as:

Whilst outsourcing translation projects may reduce short-term costs, the increased quality and information security risks may mean higher costs in the long-term. This paper explores the application of technology and methods that translation companies and their clients can deploy to mitigate the risks of outsourced translation and increase the probability of a successful project outcome.   The paper focuses on the use of online collaboration tools such as “virtual spaces” which can effectively support global translation projects.  Using a case study, the paper will outline key lessons learned.

11.45  Discussion

11.50 Statistical Glossing – Language Independent Analysis in Bible Translation
Jon Riding, BFBS, UK
Much of the work of the Bible translation community is aimed at developing world vernacular languages. Computer-based lexica and grammars are rarely available to assist translators. The Statistical Glossing Tool (SGT), released in 2007, helps translation consultants to review new translations and requires no input from the user other than to          indicate the general family of inflection to which the target language belongs.  This paper gives a brief description of the development of the tool, present limitations and possible future developments. The practical difficulties of building a truly language independent system are discussed as are the experiences of early adopters in the field.  The presentation will conclude with a demonstration.

12.25 Discussion

12.30 Lunch and Exhibition

14.00 Introduction by Chair: Professor Ruslan Mitkov, University of Wolverhampton, UK

14.00 Panel Discussion on the future: Moderator: Reinhard Schäler, University of Limerick, Ireland
The Future of Translation and the Computer
Predicting the future role of the computer for translation is easy: it will grow to the point where human translators will no longer be necessary; they will be replaced by specialized operators and administrators in all but a few insignificant niches. Where linguistic intervention by humans is required, social networking sites will provide the “crowd” with the necessary language background. – This rather radical view of the future of translation will be discussed by a panel of distinguished experts whose probably controversial views will be discussed with the participants of this year’s 30th ASLIB conference.

15.00  Tea

15.20 Developing a 12 Week e-Learning Course on Software Localisation at Imperial College London: Design and Insight From Delegate Feedback
Mark Shuttleworth and Daniela Ford, Imperial College London, UK
The twelve week stand alone e-learning course provides Imperial College with an opportunity to make the course available to all of its translation students (irrespective of their language combination) and to use it as an adjunct to the existing teaching.  In addition, the course will be made available to external delegates also.  The aim of this paper is to present an overview of the practical design, implementation and testing steps for the new e-learning course on software localization and analysis of the feedback received.  It is hoped that this will provide understanding and some useful pointers on how others can design such a course and to show how both academia and the localization industry may be able to draw benefits from this.

15.55  Discussion

16.00 Beyond Terms: Multi-Word Units in MultiTerm Extract
María Fernández Parra and Pius ten Hacken, Swansea University, WalesMulti-word units are expressions whose form, syntax, or meaning have non- compositional as well as compositional components.  These units are frequent in texts, including in the type of text that is translated with the help of CAT tools.  The     non-compositional nature of multi-word units makes it desirable for CAT tools to provide support in their recognition and translation.  In SDL Trados 2007, there are two databases for the recording of previous translations, the Translation Memory for segments and the MultiTerm termbase for terms.  SDL Trados 2007 offers a number of tools to support the translation of terms, including MultiTerm Extract.  If we consider the termbase as the place to store multi-word units, it is interesting to find out whether these tools can be used successfully with multi-word units instead of actual terms.  A number of experiments were carried out on the recognition of multi-word units by MultiTerm Extract.  This paper proposes a set of recommendations for translators on how to use MultiTerm Extract for the recognition of multi-word units.  By analysing the type of problems that still remain, avenues of development for the software will also be suggested.

16.35 Discussion

16.40 Terminological Resources for Translators: A Survey       
Isabel Durán-Muñoz, University of Málaga, Spain
This project seeks to identify the real needs of translators with regards to
terminology.  More specifically, this paper will present the results of a recent survey in which translators were asked which terminological resources they currently use and what resources they would ideally like to use.    The survey will identify the needs and will report on the expectations and desires of translators, thereby illustrating what their "ideal" terminological resource would be.  Unfortunately, the majority of resources currently available are of little use to translators, which is why many must resort to the creation of their own terminological resources either from parallel corpora or from existing translations.   However, it is hoped that the identification of the user's needs with regard to terminology could lead to useful resource development projects in
the future.

17.15 Discussion

17.20 Close of Conference

Please note: Aslib reserves the right to make changes to the publicised programme without prior notice.

---

back to the top of the page

Book your place now