Chris Dyer's Homepage
I am a PhD candidate in the Department of Linguistics at the University of Maryland at College Park working with Philip Resnik. My primary focus is computational linguists (my interests include machine translation, unsupervised learning, and computational phonology and morphology), but problems from a variety of other areas of linguistics and computer science interest me as well.
I participated in the 2006 Summer Workshop at Johns Hopkins which produced the Moses open source toolkit for statistical machine translation.
I'm working with Jimmy Lin in CLIS on using large networks of computers (provided by IBM and Google-- thanks!) to process large amounts of natural language data efficiently. Read the hype! We had a tutorial about it NAACL 2009, and now we're writing a book about the experience.
For a couple months in 2009, I worked in the ICCS, part of the School of Informatics at Edinburgh University, where my esteemed colleagues Phil Blunsom, Miles Osborne, and Philipp Koehn provided structure and guidance, and beer. While there, I decided that I like Monte Carlo simulations, even if they aren't as glamourous as the real Monte Carlo.
I spent the summer of 2008 at Google, working on machine translation. We then wrote a paper about my inaccurately named "20% project".
For fun, I play cello. I also like to think about the cognitive aspects of music theory as well as the relationship of language to music.
Here's a somewhat up-to-date CV.
I recently discovered I have an Erdös number of 4.
Refereed Publications/Presentations
P. Blunsom, T. Cohn, C. Dyer, and M. Osborne. A Gibbs sampler for phrasal synchronous grammar induction. In Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), Singapore, August 2009.
S. Kumar, W. Macherey, C. Dyer, and F. Och. Efficient minimum error rate training and minimum Bayes-risk decoding for translation hypergraphs and lattices. In Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), Singapore, August 2009.
C. Dyer. Using a maximum entropy model to build segmentation lattices for MT. In Proceedings of NAACL HLT 2009, Boulder, Colorado, June 2009. [slides]
A. Arun, C. Dyer, P. Blunsom, B. Haddow, A. Lopez, and P. Koehn. Monte Carlo inference and maximization for phrase-based translation. In Proceedings of CoNLL, Boulder, Colorado, June 2009.
C. Dyer, H. Setiawan, Y. Marton, and P. Resnik. The University of Maryland Statistical Machine Translation System for the Third Workshop on Machine Translation. In Proceedings of the EACL-2009 Workshop on Statistical Machine Translation (WMT-2009), Athens, Greece, March 2009.
Z. Li, C. Callison-Burch, C. Dyer, J. Ganitkevitch, S. Khudanpur, L. Schwartz, W. Thornton, J. Weese, and O. Zaidan. 2009. Joshua: Open Source Toolkit for Parsing-based Machine Translation. In EACL 2009 Fourth Workshop on Statistical Machine Translation, Athens, Greece, March 2009.
C. Dyer, A. Cordova, A. Mont, and J. Lin. Fast, Cheap, and Easy: Construction of Statistical Machine Translation Models with MapReduce. In Proceedings of the ACL-2008 Workshop on Statistical Machine Translation (WMT-2008), Columbus, Ohio, July 2008.
C. Dyer, S. Muresan, and P. Resnik. Generalizing Word Lattice Translation. In Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), Columbus, Ohio, July 2008.
C. Dyer. The University of Maryland Translation System for IWSLT 2007. In Proceedings of the International Workshop on Spoken Language Translation (IWSLT) 2007, Trento, Italy, October 2007.
C. Dyer. The 'noiser channel': Translation from morphologically complex languages. In Proceedings of the ACL-2007 Workshop on Statistical Machine Translation (WMT-2007), Prague, Czech Republic, June 2007.
P. Koehn, H. Hoang, A. Birch, C. Callison-Burch, M. Federico, N. Bertoldi, B. Cowan, W. Shen, C. Moran, R. Zens, C. Dyer, O. Bojar, A. Constantin and E. Herbst. Moses: Open Source Toolkit for Statistical Machine Translation. In Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), Companion Volume, Prague, Czech Republic, June 2007.
A. Omaki, C. Dyer, S. Malhotra, J. Sprouse, J. Lidz, & C. Phillips. The time-course of anaphoric processing and syntactic reconstruction. Presented at CUNY 2007, San Diego, March 2007.
Other talks
C. Dyer. Improving machine translation by propagating uncertainty. Presented at IBM TJ Watson Research Center, Yorktown Heights, NY, August 2009.
J. Lin and C. Dyer. Data-Intensive Text Processing with MapReduce. Tutorial presented at the North American Chapter of the Association for Computational Linguistics - Human Language Technologies (NAACL HLT) 2009 Conference.
C. Dyer. Decoder guided backoff: using word lattices to improve translation from morphologically complex languages. Presented at the MT Marathon, Edinburgh University, April 2007.
C. Dyer and P. Resnik. Confusion network decoding in Hiero. Presented by P. Resnik at GALE 2007 PI Meeting, San Fransisco, March 2007.
C. Dyer and J. Lin. Computing in the Clouds: Applications of MapReduce in "Web-Scale" Information Processing. Presented at the University of Maryland CLIP Colloquium, October 17, 2007.
C. Dyer. Decoder guided backoff: using word lattices to improve translation from morphologically complex languages. Presented at the MT Marathon, Edinburgh University, April 2007.
C. Dyer and P. Resnik. Confusion network decoding in Hiero. Presented by P. Resnik at GALE 2007 PI Meeting, San Fransisco, March 2007.
C. Dyer. The Syntax of Tonal Music, Syntax Lunch Talk, University of Maryland, October 2006.
Software and data
- libpopt is a small, fast C++ library that implements a few gradient-based optimization algorithms (LBFGS, stochastic gradient descent, Rprop) where the optimizer's state can be persisted between function evaluations. This makes it possible to perform successive optimization iterations in different processes or on different machines (useful with MapReduce).
- In a fit of rage and frustration that no such tool existed, I wrote a very small, very fast Chinese word segmenter that's in C++.
- GIZA++ and MkCls, "fixed" to compile with recent GCCs.
- A Java implementation of the Buckwalter Morphological Analyzer v1 (GPL License, data (c) 2001 by the LDC).
- Chinese place names (Chinese (utf-8) | English) which were mined from some huge lists at Wikipedia (URL list).
