Information
|
|
|
Research
My research involves combining statistical methods with linguistic knowledge to improve the quality of automatic translation between human languages. I work with Alon Lavie in the AVENUE machine translation group where I focus on automatic evaluation and tuning of MT systems. I currently develop the Meteor automatic MT evaluation metric.
Software
Meteor Automatic Machine Translation Evaluation System
[webpage] [github] [GNU LGPL]
Alignment-based MT evaluation metric with extended support for several target languages. The current version includes paraphrase tables for five languages as well as tools for visualizing alignments and score distributions.
Parex Paraphrase Extractor
[github] [download]
Simple tool for extracting paraphrases from parallel corpora using phrase tables, includes support for language-independent filtering. Parex was used to build the Meteor paraphrase tables.
MT-Diff
[Now included with Meteor]
Measure segment level differences in BLEU and Meteor between two sets of translation hypotheses, report number of segments different and magnitudes of differences.
Select Publications
M. Denkowski, G. Hanneman, A. Lavie,
"The CMU-Avenue French-English Translation System",
Proceedings of the NAACL 2012 Workshop on Statistical Machine Translation
[PDF]
[bib]
•★ Win and constrained win in shared translation task
M. Denkowski and A. Lavie,
"Meteor 1.3 Automatic Metric for Reliable Optimization and Evaluation of
Machine Translation Systems",
Proceedings of the EMNLP 2011 Workshop on Statistical Machine Translation
[PDF]
[bib]
• Tunable metrics task win in WMT11
★ Segment level win (or tied) for into and out-of English tasks in WMT12
M. Denkowski and A. Lavie, "Choosing the Right Evaluation for Machine Translation: an Examination of Annotator and Automatic Metric Performance on Human Judgment Tasks", Proceedings of AMTA, 2010 [PDF] [slides]
M. Denkowski and A. Lavie, "METEOR-NEXT and the METEOR Paraphrase Tables: Improved Evaluation Support For Five Target Languages", Proceedings of the ACL 2010 Joint Workshop on Statistical Machine Translation and Metrics MATR [PDF] [bib]
M. Denkowski, H. Al-Haj, A. Lavie, "Turker-Assisted Paraphrasing for English-Arabic Machine Translation", Proceedings of the NAACL HLT 2010 Workshop on Creating Speech and Language Data With Amazon’s Mechanical Turk, 2010 [PDF] [bib]
M. Denkowski and A. Lavie, "Exploring Normalization Techniques for Human Judgments of Machine Translation Adequacy Collected Using Amazon Mechanical Turk", Proceedings of the NAACL HLT 2010 Workshop on Creating Speech and Language Data With Amazon’s Mechanical Turk, 2010 [PDF] [bib]
M. Denkowski and A. Lavie, "Extending the METEOR Machine Translation Evaluation Metric to the Phrase Level", Proceedings of NAACL/HLT, 2010 [PDF] [bib]
A. Lavie and M. Denkowski, "The METEOR Metric for Automatic Evaluation of Machine Translation", Machine Translation, 2010 [PDF]
Reports
M. Denkowski and A. Lavie, "METEOR-Tuned Phrase-Based SMT: CMU French-English and Haitian-English Systems for WMT 2011", Technical Report CMU-LTI-11-011, Language Technologies Institute, Carnegie Mellon University, 2011 [PDF] [bib]
M. Denkowski, "A Survey of Techniques for Unsupervised Word Sense Induction", Language & Statistics II Literature Review, Fall 2009 [PDF]