Information

Michael Denkowski
Michael Denkowski
PhD Candidate

Language Technologies Institute
Carnegie Mellon University
5000 Forbes Avenue
Pittsburgh, PA 15213

Google map
Office: Gates Hillman Center 5709
Email:

Research

My research focuses on improving the quality of statistical translation between human languages. I work with Alon Lavie in the AVENUE machine translation group on topics including MT for human translators (post-editing), automatic metrics for system optimization and evaluation, and building large scale systems. See below for a list of software I've worked on or take a look at my github page.

Software

Realtime Adaptive Machine Translation for Human Translators (cdec-realtime)
[tutorial] [github] [Apache License]
Machine translation that learns from human feedback. Use incremental post-editing data to update translation models, language models, and feature weights with minimal approximation.

Meteor Automatic Machine Translation Evaluation System
[webpage] [github] [GNU LGPL]
Alignment-based MT evaluation metric with extended support for any target language as well as tools for visualizing alignments and score distributions.

Quality Estimation-inspired Parallel Data Cleaning
[README] [github] [GNU LGPL]
Select the most reliable data from large parallel corpora using fast, simple features from MT quality estimation. Qe-clean was used to clean parallel data for our WMT systems in 2012 and 2013.

TransCenter Web-Based Translation Research Suite
[webpage] [github] [GNU LGPL]
TransCenter allows you to efficiently collect and analyze human translation data over the web. TransCenter server is easy to deploy and maintain. The web-based translation editor allows translators to work on translation tasks from any computer with an Internet connection. All user activity is logged so you can view detailed translation reports.

Parex Paraphrase Extractor
[github] [GNU LGPL]
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.

Select Publications and Talks

Michael Denkowski, "Learning from Post-Editing: Real Time Model Adaptation for Machine Translation", Invited talk at the AMTA 2014 Workshop on Interactive and Adaptive Machine Translation, 2014
[slides]

Michael Denkowski; Alon Lavie, "Meteor Universal: Language Specific Translation Evaluation for Any Target Language", Proceedings of the ACL 2014 Workshop on Statistical Machine Translation, 2014
[PDF] [bib] [poster] [tutorial]
Build a version of Meteor for any language!

Michael Denkowski; Alon Lavie; Isabel Lacruz; Chris Dyer, "Real Time Adaptive Machine Translation for Post-Editing with cdec and TransCenter" Proceedings of the EACL 2014 Workshop on Humans and Computer-assisted Translation, 2014
[PDF] [bib] [poster] [tutorial] [cdec Realtime] [TransCenter]

Michael Denkowski; Chris Dyer; Alon Lavie; "Learning from Post-Editing: Online Model Adaptation for Statistical Machine Translation" Proceedings of EACL, 2014
[PDF] [bib] [poster] [tutorial] [github]

Olga Beregovaya; Alon Lavie; David Clarke; Michael Denkowski, "Analyzing and Predicting MT Utility and Post-Editing Productivity in Enterprise-scale Translation Projects" Proceedings of Machine Translation Summit XIV, 2013
[Abstract] [Slides]

Michael Denkowski, "Machine Translation for Human Translators" Ph.D. Thesis Proposal, Carnegie Mellon University, May 2013. Committee: Alon Lavie; Chris Dyer; Jaime Carbonell; Gregory Shreve
[PDF] [bib] [slides]
Build your own adaptive translation system!

Michael Denkowski; Alon Lavie, "Challenges in Predicting Machine Translation Utility for Human Post-Editors" Proceedings of AMTA, 2012
[PDF] [bib] [slides]

Michael Denkowski; Greg Hanneman; Alon Lavie, "The CMU-Avenue French-English Translation System", Proceedings of the NAACL 2012 Workshop on Statistical Machine Translation, 2012
[PDF] [bib]

Michael Denkowski; Alon 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, 2011
[PDF] [bib]

Michael Denkowski; Alon 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]

Michael Denkowski; Alon 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, 2010
[PDF] [bib]

Michael Denkowski; Hassan Al-Haj; Alon 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]

Michael Denkowski; Alon 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]

Michael Denkowski; Alon Lavie, "Extending the METEOR Machine Translation Evaluation Metric to the Phrase Level", Proceedings of NAACL/HLT, 2010
[PDF] [bib]

Book Chapters

Lavie; Agarwal; Denkowski; Snover; Madnani; Dorr; Schwartz; Habash; Kahn; Ostendorf; Roark; Kulick; Marcus; Pado; Galley; Manning; "Searching for Better Automatic MT Metrics", Handbook of Natural Language Processing and Machine Translation, 2011

Reports

Michael Denkowski; Alon Lavie, "TransCenter: Web-Based Translation Research Suite", AMTA 2012 Workshop on Post-Editing Technology and Practice Demo Session, 2012
[PDF] [bib]

Michael Denkowski; Alon 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]

Michael Denkowski, "A Survey of Techniques for Unsupervised Word Sense Induction", Language & Statistics II Literature Review, Fall 2009
[PDF]