Preliminary Program of MT Marathon 2009
Monday Jan 26 - Friday Jan 30, 2009
Apart from lectures, talks and lab sessions, there will be group projects (aka Open Source Convention). For more details see:
MT Marathon 2009 Wiki
read-only access: username: mtm, password: mtm
Lectures, Talks, Labs
Registration starts at 8.30 in front of room S1.
Morning Lecture | Research Talks | Afternoon Lab Session | Evening | |
---|---|---|---|---|
9.00-10.30, room S1 | 11.00-12.30, room S1 | 14.00-17.00, lab SU2 | 17.00-?, room S1 | |
Monday |
General welcome and local info (Philipp Koehn and Jan Hajič, 9.00-9.10) Introduction to MT and MT Evaluation (Adam Lopez) |
|
Manual judgement of MT quality | |
Tuesday | Word Alignment (Barry Haddow) | PostCAT, apertium-cy, MBMT | Implementing IBM model 1 | Projects: Update |
Wednesday | Phrase-Based Models and Decoding (Chris Dyer) | Joshua, SAMT, MERT+ | Installing and running Moses (Hieu Hoang and Josh Schroeder) | |
Thursday | TectoMT: Processing Trees (Zdeněk ®abokrtský and others) | RIA, Sub-Tree Aligner | TectoMT hands-on experience: Installation and tutorial (Jana Kravalová) | Projects: Update |
Friday | Richer Models and Optimization (Philipp Koehn) |
Projects: Final Short Presentations
Panel: Future of MT Marathon |
Using factored models and MERT in Moses (Hieu Hoang, Barry Haddow, Abhishek Arun). |
Accepted Contributions, Research Talks
The following contributions will be presented during late mornings:
- apertium-cy: F. M. Tyers, K. Donnelly: apertium-cy - a collaboratively-developed free RBMT system for Welsh to English
- Joshua: Z. Li, C. Callison-Burch, W. Thornton, S. Khudanpur: Joshua: an Open-source Decoder for Parsing-based Machine Translation
- MBMT: A. van den Bosch and P. Berck: Memory-Based Machine Translation and Language Modeling
- MERT+: N. Bertoldi, B. Haddow, J.-B. Fouet: Improved Minimum Error Rate Training in Moses
- PostCAT: J. Graça, K. Ganchev, B. Taskar: PostCAT - Posterior Constrained Alignment Toolkit
- RIA: Y. Graham, J. van Genabith: An Open Source Rule Induction Tool for Transfer-Based SMT
- SAMT: A. Venugopal, A. Zollmann: Grammar based statistical MT on Hadoop. An end-to-end toolkit for large scale PSCFG based statistical machine translation
- Sub-Tree Aligner: V. Zhechev: Unsupervised Generation of Parallel Treebanks through Sub-Tree Alignment
- Z-MERT: O. Zaidan: Z-MERT: A Fully Configurable Open Source Tool for Minimum Error Rate Training of Machine Translation Systems (There is no presentation for this paper.)
Research talk presentations should be 20 to 25 minutes long with additional 5 minutes for a discussion.
Past MT Marathons
The program structure follows previous marathons:
- First MT Marathon, Edinburgh 2007
- Second MT Marathon, Berlin/Wandlitz 2008; More details here.