Statistical Multilingual Analysis for Retrieval and Translation - Barcelona 2009

A joint event of SMART project - Pascal Network

Co-organizers: Marco Turchi, Nello Cristianini, Xavier Carreras, Tijl de Bie

Barcelona May 13, 2009

Venue: Aula Teleensenyament (Tele-teaching room) in building B3
of the Campus Nord of the UPC

The aim of this workshop is to disseminate scientific results produced by the SMART project to the larger technical and scientific community working on Statistical Machine Translation. To facilitate this inter-exchange, it will be co-located with

EAMT 2009 - 13th Annual Conference of the European Association for Machine Translation

That will be held May 14-15, 2009
Universitat Politècnica de Catalunya, Barcelona, SPAIN
Conference web site: http://www.talp.cat/eamt09
Workshops page: http://www.talp.cat/eamt09/index.php/associated-workshops

Scientific Programme

Morning

9.30 - 10.00 Welcome, Nicola Cancedda (Xerox Research Centre Europe)

10.00 - 11.00 Invited Talk: "Empirical Machine Translation and its Evaluation" Jesus Gimenez, UPC

11.00 - 11.30 Coffee

11.30 - 12.00 - "Online learning for CAT applications" Nicolo Cesa-Bianchi (University of Milan)

12.00 - 12.30 - "Sinuhe -- Statistical Machine Translation with a Globally Trained
Conditional Exponential Family Translation Model" - Matti T Kaariainen (University of Helsinki)

12.30 - 1300 - Large scale, maximum margin regression based, structural learning approach to phrase translations - Sandor Szedmak (University of Southampton)

LUNCH

Afternoon

14.00 - 14.30 "Learning to Translate: statistical and computational analysis" - Marco Turchi (University of Bristol)

14.30 - 15.00 -"Improving the performance of phrase-based statistical MT" - By: NRC

15.00 - 15.30 - "Multi-view CCA and regression CCA" includes online demo of this integrated with searchpoint integration - Blaz Fortuna (Jožef Stefan Institute)

Coffee

16.00 - 16.30 - "Large-Margin Structured Prediction via Linear Programming" - Zhuoran Wang (UCL)

16.30 - 17.00 - "Confidence Estimation for Machine Translation" - Lucia Specia (XRCE)

17.00 Closing Remarks

ABOUT THIS WORKSHOP

A joint event of SMART project - Pascal Network

SMART (Statistical Multilingual Analysis for Retrieval and Translation) is a 3-year "Specific Target Research Project" (STReP) funded by the European Commission. SMART is an attempt to address different problems of Machine Translation and Cross-Language Information Retrieval and other shortcomings by the methods of modern Statistical Learning.

In the first two years of the project, the scientific focus has been on developing new and more effective statistical approaches while ensuring that existing know-how is duly taken into account. By was done by bringing together leading research institutions in Statistical Learning, Machine Translation and Textual Information Access.

The aim of this workshop is to disseminate scientific results and share experiences produced by the SMART project to the larger technical and scientific community. The SMART consortium considers this workshop to be a great opportunity for science investigations, creating both scientific and commercial opportunities as well as technological challenges to researchers.