[Corpora-List] MT Summit X Newsletter No.1

Eiko Yamamoto eiko at nict.go.jp
Mon Mar 28 04:34:00 CEST 2005

*********** MT Summit X Newsletter No.1 (22nd March 2005)***************

The 10th Machine Translation Summit [MT Summit X]


Main Conference: September 13-15, 2005
Tutorials & Workshops: September 12 and 16, 2005

Hilton Phuket Arcadia Resort and Spa
This issue contains

[1] Messages from Conference Chair/Local Organising Chair
[2] Topics on the Main Program
[3] Extension (referring to the deadline)
[4] Workshops
[5] Tutorials
[6] Social Events
[7] Registration Fees
[8] Call for Exhibitors

[1] Messages from Conference Chair/Local Organising Chair

As you are aware, Thailand was hit very hard by the recent tsunami,
prompting the MT Summit X Steering Committee to carefully reconsider
the venue. However, all the news from the local organizing committee
and the conference venue (Hilton Phuket Arcadia Resort and Spa) has
been very positive about having the Summit as we planned.
The local organizing committee and the steering committee have
received many supportive messages from members of AMTA, EAMT and
AAMT. Many suggested that the Summit in Phuket would boost the morale
of the local people and thus contribute to local recovery efforts. The
suggestion is fully confirmed by the local organizing committee.
The steering committee, therefore, has decided to hold with our
decision of having the MT Summit X in Phuket from September 12 to 16,
2005. We are considering of ways of showing our sympathy and support
to the local people. Some of our plans are in this newsletter and more
will follow.
We are looking forward to seeing you all at the MT Summit X in

Junichi Tsujii, President of AAMT
Conference Chair, MT Summit X
February 19, 2005

On December 26, 2004, a massive number of lives were destroyed by the
tsunami. Six provinces in Southern Thailand facing the Andaman Sea,
namely Phuket, Phang-nga, Ranong, Krabi, Trang and Satul were hit by
the disaster. In only a matter of hours thousands of people lost their
families, houses, schools and businesses. Phuket Island was reported
to have had the third largest amount of damage of about 2,000 missing,
dead and injured. We have all felt emotions stirred by the images of
those affected by the disaster.
Despite the large amount of damage reported, the six provinces are
gradually recovering by the intensive assistance from the cooperation
among national and international organizations. Phuket has the biggest
economy among the six affected provinces.
The faster the Phuket Island economy is restored, the more quickly the
other areas will begin to recover. From the perspective of
re-establishment of the resort town, Thai government is now working
hard to bring back the Phuket attractions as well as establishing an
appropriate warning system to secure the area.
The gathering for MT Summit X in Phuket is greatly welcomed not only
to relieve the local people of the tragedy but also to e[6] Social
Eventsxplore the technological possibility of utilization of MT for
communication in emergencies. In the recent tsunami disaster, ICT
(Information and Communications Technology) played an important role
in the urgent needs of database construction to support the management
and searching of the disaster information via the Internet. The
collaborative work using web technology has proven its strength in
responding to the sudden growth of the information flow.
Under the well-established infrastructure of the Internet, the coming
MT Summit X will be an appropriate place for seriously discussing the
MT technology for emergency communication. Language is still a great
barrier. We lack off-the-shelf technology to provide the just-in-time
service. This could be one of the challenging topics of the
contemporary MT research.
On behalf of the local organizing committee, I would like to express
my warm welcome to the Summit participants to Phuket. One of your
contributions will be in part to restore normal lives and joyfulness
to the local people. Once again, I would like to call for a new
challenging research of MT technology for emergency communication that
includes cross language, cross culture or language independent
communication. In addition, Hilton Phuket Arcadia Resort and Spa, the
selected location for the MT Summit X venue, is fully operating
without any significant damage. It currently serves many international
meetings especially on topics related to the tsunami.

Virach Sornlertlamvanich
Local Organizing Chair, MT Summit X
February 3, 2005

[2] Topics on the Main Program

(a) Invited Speakers

We will invite distinguished speakers to this summit.
Tentative list of invited speakers include;

1) Prof. Makoto Nagao, President of NICT (Tokyo)
2) Prof. Dr.-Ing. Hermann Ney, RWTH (Aachen)
"Statistical Machine Translation."
3) Dr. Heyan Huang, CCLIE-CAS and Huajian MT systems (Beijing)
"MT reseach center and vendor in China"
4) Dr. Liu Weiquan, Beijing Network & Multimedia Laboratory (Beijing)
5) Dr. Gregor Thurmair, Linguatec (Germany)
"Improving Machine Translation Quality"
6) Prof. Toru Ishida, Kyoto University (Kyoto)
"Inter-cultural collaboration using MT"
7) Prof. Bo Xu, CASIA (Beijing)
"Speech Translation"

More will be added in the following newsletters.

(b) Panel

We will have several panels to encourage undersatndings among

(1) User panel, which will follow up the user workshop in the summit.
(2) "Bounds of SMT"
MT researchers from different parspectives, such as SMT, EBMT and
RBMT, will participate this panel and discuss how their approach
can be the best way to develop useful MT systems.
(3) "Lesson learnt from MMT project" in which participants of MMT
project (Multi-lingual Machine Translation Project by Japanese
government from 1987 to 1995) will discuss their experiences and
what follows the project.

(c) Organized Session

We will have several organized sessions which cover specific area of
MT developments.

(1) MT evaluation
(2) MT for emergency communication
(3) Asian language resources and processings for MT
[3] Extension (referring to the deadline)

Considering several factors relating to the process of making our
proceedings, we decided toextend the submission due. Current schedule
is as follows;
Submmission due: 13th of May.
Notification to the authors: 24th of June.
Camera ready due: 22nd of July.
Conference: 12th-16th of September.
[4] Workshops:
MT SUMMIT X will have 5 workshops as follows.

(1) 2005/09/16, Open-Source Machine Translation,
Prof. Mikel L. Forcada (Universitat d'Alacant, Spain)

(2) 2005/09/16, Example-Based Machine Translation,
Dr. Michael Carl (IAI/Universitat des Saarlandes)
Prof. Andy Way (School of Computing/Dublin City University)

Following the successful EBMT workshop in 2001
<http://www.compapp.dcu.ie/%7Eaway/EBMT.html> at the MT Summit VIII
<http://www.eamt.org/summitVIII> in Santiago de Compostela, Spain,
is the second workshop of its kind. EBMT approaches have evolved since
2001, with some even reaching commercial maturity. Furthermore, a
of new and innovative approaches have been seen. Their common
characteristics are to exploit and integrate a number of knowledge
resources, such as linguistics and statistics (or symbolic and
numerical techniques) for integration into a single framework. In this
way, rule-based morphological, syntactic and/or semantic information
is combined with knowledge extracted from bi- or monolingual texts
which is then re-used in the translation process. We seek to assemble
approaches to machine translation which go beyond the purely
statistical and/or rule-based paradigms.

(3) 2005/09/16, Patent Translation,
Prof. Shoichi Yokoyama (Yamagata University)

Patent information is one of the major application areas of machine
translation. The workshop aims to foster research and development of
the technology for patent translation by providing a forum in which
researchers and practitioners can exchange their ideas, approaches,
perspectives, and experiences from their work in progress. The
workshop will consist of invited talk(s), presentation of submitted
papers, and one panel discussion.

Topics of interests include:
- Analysis and classification of patent documents
- MT and translation aids for patent documents
- Contrastive studies for multilingual patent documents
- Language resources for patent translation
- Information extraction from patent documents
- Evaluation techniques for patent translation
- Multilingual patent classification and retrieval

Important Dates:
Paper submission deadline: 30 May
Notification of acceptance: 30 June
Camera ready submission deadline: 20 July

The workshop homepage will be set up soon.

(4) 2005/09/16, Interlinguas and Interlingual Approaches to MT: Spoken
Language Translation,
Prof. David Farwell (CRL-NMSU)

This 8th Workshop on Interlinguas brings together specialists in MT
and NLP to work out practical representation systems for a range of
pragmatic and semantic phenomena related to speech-to-speech
translation systems and the translation of spoken dialogue.

Papers are invited on the interlingual representation and/or
processing of:
- Dialogue structure, discourse structure and rhetorical structure,
- conversation analysis or conversational structure (turn-taking),
- speech acts (including pauses),
- conversational implicature,
- reference resolution and/or the recovery of ellipted information,
- extra-linguistic context and deictic references,
- interlocutor (speaker-addressee) or interpersonal relationships
(e.g., politeness),
- disfluencies and/or erroneous speech recognition output,
- other discourse or dialogue related phenomena.

Visit http://crl.nmsu.edu/Events/FWOI/EighthWorkshop/index.html for
the Call for Papers and other details.

(5) 2005/09/12, Semantic Web Technologies for Machine Translation,
Dr. Cristina Vertan (University of Hamburg)
Prof. Dr. Walther v. Hahn (University of Hamburg)
Dr. Vladislav Kubon (Charles University Prague)

The goal of the workshop is twofold:
- to discuss the implications of semantic web-technologies for machine
translation, namely on example based and knowledge-based machine
- to contrast the two main technologies of Semantic Web: topic maps
and RDFS, in machine translation of on-line texts.

We welcome original papers related (but not limited) to following
- semantic web annotations for multilingual corpora
- use of semantic web annotations for corpus based machine translation
- integration of semantic information in example based machine
- use of semantic web ontologies for machine translation
- semantic web and on-line translation tools
- integration of semantic web technologies in CAT tools.

We also encourage demonstrations of developed tools. Submissions for a
demonstration session should include a 2 page demo-note describing the
system-architecture and performance as well as technical requirements.
[5] Tutorials

MT SUMMIT X will have 3 tutorials given by leading experts in natural
language processing field.

(1) 2005/09/12 AM, Statistical Machine Translation: Foundations and
Advances, Franz Och (Google)

The use of statistical techniques in machine translation has recently
attracted a considerable amount of interest. This tutorial will
discuss the theoretical foundations of statistical machine translation
in detail, present an overview on the architecture of the currently
best performing systems and discuss recent advances in the field.

Selected Contents:
- What is statistical machine translation?
- Why statistical machine Ttanslation?
- Basic concepts: Model, Training, Search
- Models: source-channel model, log-linear models, ...
- Training: maximum likelihood, maximum BLEU training, ...
- Search: dynamic programming, N-Best list generation, minimum
Bayes risk decoding
- Phrase/Template-based translation models
- Results of recent machine translation evaluations
- Syntactic models

Franz Josef Och is a Research Scientist at Google, Inc.
Dr. Och is working on statistical machine translation since 1997
and participated in various European and US projects which involved
machine translation. Dr. Och has written a number of articles on
statistical machine translation and has co-authored GIZA++, a widely
used software package for training statistical alignment models.

(2) 2005/09/12 PM, Text Mining: Opportunity for Machine Translation
Technology, Tastuya Nasukawa (IBM)

This tutorial introduces text mining (TM) technology that develops
valuable knowledge from extremely large amounts of textual data by
analyzing trends and finding significant features. Although MT
(Machine Translation) and TM share much NLP technology, TM has a
distinguishing feature of taking advantage of voluminous data that
allows focusing on high frequency data extracted from the original
texts and ignoring low frequency items. In this sense, TM is designed
to compensate for failures in NLP that generate wide varieties of low
frequency items.

The presenter is hoping that this tutorial contributes new ideas
toward enhancing opportunities for MT researchers and developers by
* How TM takes advantage of NLP techniques while compensating for its
* How TM benefits from NLP techniques for MT systems
* How MT functions may be integrated into TM for handling
multilingual data for global organizations

* Overview of Text Mining
- How it works
- What it does
* Application Examples of Text Mining
- How it creates value
* Important Issues for Success with Text Mining
- What needs to be done
* Research Issues in Text Mining
- How to improve it
* Comparison of Text Mining and Machine Translation

Tetsuya Nasukawa is a senior researcher in IBM Research, Tokyo
Research Laboratory. He was involved in several machine translation
projects after joining IBM Research in 1989. He started a text mining
project to take dvantage of natural language processing technology
developed through machine translation projects, and he developed a
text analysis and knowledge mining system named IBM TAKMI. In addition
to his work within TRL, he has had several special assignments, such
as implementing MT in the National Language Support section of IBM
Japan, doing research as a research staff member at the Watson
Research Center, and supporting real-world TM implementations with IBM
Business Consulting Services.

(3) 2005/09/12 PM, Interlinguas and Semantic Roles, Mike Dillinger
Translation), http://www.mikedillinger.com

There has been a resurgence of interest in interlinguas and semantic
roles, now focussed on knowledge processing rather than on
interlingua-based translation. New data sets such as FrameNet and
PropBank have made possible statistical approaches to mapping from
sentences to semantic roles.
Several US government projects have recently invested in semantic
roles as an approach for information extraction and knowledge
In this tutorial, we will review recent efforts to design and use
interlinguas and the new prospects for applying them to translation.

Tutorial Contents:
* Advantages and disadvantages of Interlinguas
* "Anatomy" of Interlinguas
- Design and characteristics of interlingua components
- Comparison of different interlinguas
* "Physiology" of Interlinguas
- Mapping into and out of interlinguas
- Processing interlingua statements
* Problems and prospects

Mike Dillinger, PhD is Director of Linguistics for Spoken Translation,
and an industry consultant who helps organizations make their
documentation processes significantly more effective. He is also a
member of the Board of Directors of AMTA and Assistant Director of the
Translation Automation Users Society (TAUS). At Spoken he supervises
the design and creation of semantic knowledge bases for interactive
translation. Previously, he was Director of Linguistic Development for
two MT companies and he consulted for the Universal Networking
Language Project of the United Nations University in Tokyo, where he
wrote specifications for the Universal Networking Language interlingua
as part of a global effort to implement distributed, interlingua-based
machine translation. He has taught semantics and other disciplines at
more than a dozen universities in several countries over the last
twenty years, and has been a visiting researcher and/or consultant on
four continents.
[6] Social Events

We will have a reception and a banquet during the summit.
Followings are tentative plan for our social events.

(a) Reception
Date: September 13, 2005
Time: 18:30 - 21:00
Place: Hilton Pool Side
Style: Elegant Buffet, Classical Thai Dance
All participants are invited.

(b) Banquet
Date: September 14, 2005
Time: 19:00 - 21:30
Place: Baan Klung Jinda Restaurant, in the heart of Phuket
Style: Siamese Cuisine & International luxurious dinner set

Explore the Pearl of Andaman, and the Sino-Portuguese and
Mediterranean style of Old Phuket along the fantastic road to
Baan Klung Jinda.

Fees will be announced later.
[7] Registration Fees
The registration fees are as follows.

Before August 31
- Membership Fee: US$400
- Non-Membership Fee: US$440
- Student: US$100

After September 1 and on site registration
- Membership Fee: US$450
- Non-Membership Fee: US$500
- Student: US$100

Online registration is provided on the MT Summit website. Please fill
out the registration form and submit before the deadline, the end
of August. For the late registration, on-site registration is also

#Note: The reduced registration fee (same as that for students) is
considerable for non-student participants from the countries that
suffer low exchange rates of local currencies.
The detailed information for application will be available soon in
our Website (http://www.tcllab.org/Pages/mtsummit.html)
[8] Call for Exhibitors

We will provide industrial sponsors with exhibition booths with full
Internet access at the conference site. Not only companies but also
research institutes and universities are encouraged to have exhibits.
For details, please contact exhibit at aamt.info.

More information about the Corpora-archive mailing list