Kevin Duh

Signal, Speech, and Language Interpretation Lab Department of Electrical Engineering University of Washington Seattle, WA 98195 Email: kevinduh (at) u.washington.edu My Resume (pdf)

Graduate Research Assistant

• Jan 2004 - present: PhD program, University of Washington, Electrical Engineering (expected graduation: Spring 2009)
• May 2003: BS Degree, Rice University, Electrical & Computer Engineering

Research Interests

• Machine Learning and Statistics (semi-supervised learning, structured prediction, kernel methods, graphical models, machine learning for signal processing)
• Natural Language Processing (machine translation, NLP for resource-poor languages)
• Information Retrieval (web search, learning to rank)
• Speech Recognition (language modeling, domain adaptation)

Previously I have worked at several fun places, including Microsoft Research, NTT (Japan), and MIT Lincoln Labs.

Publications

1. K. Duh and K. Kirchhoff, Learning to Rank with Partially-Labeled Data, ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR 2008)
2. K. Duh and K. Kirchhoff, Beyond Log-linear Models: Boosted Minimum Error Rate Training for N-best Re-ranking, ACL 2008 (Short Paper)
3. K. Duh, Ranking vs. Regression in Machine Translation Evaluation, Proc. of ACL 2008 Workshop on Statistical Machine Translation (Shared Evaluation Task), Columbus, Ohio, 2008.

4. A. Axelrod, M. Yang, K. Duh, K. Kirchhoff, The UW Machine Translation System for ACL WMT 2008, Proc. of ACL 2008 Workshop on Statistical Machine Translation (Shared Translation Task), Columbus, Ohio, 2008

5. K. Kirchhoff, J. Bilmes, and K. Duh, Factored Language Models - a Tutorial, UW-EE Technical Report, 2008-0004, Dept. of Electrical Eng., Univ. of Washington, Feb 2008.

6. K. Kirchhoff, K. Duh, and C. Lim, The University of Washington Machine Translation System for IWSLT 2006, Proc. of the International Workshop on Spoken Language Translation, Kyoto, Japan, 2006. [slides]
7. K. Duh and K. Kirchhoff, Lexicon Acquisition for Dialectal Arabic Using Transductive Learning, Proc. of EMNLP (Empirical Methods in Natural Language Processing), Sydney, Australia, 2006.
8. K. Kirchhoff, M. Yang, and K. Duh, Statistical Machine Translation of Parliamentary Proceedings Using Morpho-Syntactic Knowledge, Proc. of the TC-STAR Workshop on Speech to Speech Translation, Barcelona, Spain, 2006.
9. S. Corston-Oliver, A. Aue, K. Duh, E. Ringger, Multilingual Dependency Parsing using Bayes Point Machines, Proc. of NAACL-HLT 2006, New York.
   (Slides of the talk I gave at Academia Sinica are available. The slides discusses parsing as structured classification and learning by Bayes Point Machines in more depth, but please refer to the paper for updated results.)
10. K. Kirchhoff, D. Vergyri, J. Bilmes, K. Duh, and A. Stolcke, Morphology-based Language Modeling for Arabic Speech Recognition, Computer Speech and Language 20(4), 2006, pp.589-608
11. K. Duh and K. Kirchhoff, Structured Multi-label Transductive Learning: a Case Study in Lexicon Acquisition, NIPS 2005, Advances in Structured Learning for Text and Speech Processing Workshop, Whistler, Canada, Dec 2005.
12. K. Duh and K. Kirchhoff, POS Tagging of Dialectal Arabic: A Minimally Supervised Approach, 43rd Annual Meeting of the Assoc. for Computational Linguistics (ACL2005), Workshop on Computational Approaches to Semitic Languages, Ann Arbor, Michigan, USA, June 2005.
13. K. Duh, Jointly Labeling Multiple Sequences: A Factorial HMM Approach, 43rd Annual Meeting of the Assoc. for Computational Linguistics (ACL 2005), Student Research Workshop, Ann Arbor, Michigan, USA, June 2005.
14. C. Bartels, K. Duh, J. Bilmes, K. Kirchhoff, S. King, Genetic Triangulation of Graphical Models for Speech and Language Processing, Proceedings of Eurospeech/Interspeech 2005, Lisboa, Portugal, September 2005.
15. D. Vergyri, K. Kirchhoff, K. Duh, A. Stolcke, Morphology-Based Language Modeling for Arabic Speech Recognition, International Conference on Spoken Language Processing (ICSLP/Interspeech 2004), October 2004.
16. K. Duh and K. Kirchhoff, Automatic Learning of Language Model Structure, Proceedings of the 20th International Conference on Computational Linguistics (COLING-2004), Geneva, Switzerland, August 2004.
17. K. Duh and K. Kirchhoff, Automatic Learning of Language Model Structure (Extended version of COLING-2004 paper), UW-EE Technical Report 2004-0014, Dept. of Electrical Engineering, University of Washington, April 2004.
18. K. Duh, Least Squares Estimation of GPS L1-L2 Satellite and Receiver Bias, MIT Lincoln Laboratory Project Report, Nov 2003.

Downloadable Software

• Factored Language Models - now included in the SRI Language Modeling toolkit v.1.4.1 (written by Jeff Bilmes and Andreas Stolcke).
• GA-FLM - an algorithm for optimizing Factored Language Model structures
• Factored Language Model (FLM) Tutorial

Resources

• Machine Learning & Statistics
  • Introduction to Conditional Random Fields (CRFs)
  • Introduction to Causal Inference: presented with Michele Banko, at the UW Markovia seminar
  • Model Selection: lecture discussing bias-variance tradeoff, cross-validation, BIC/AIC/MDL, a bit of VC theory
• Machine Translation, Multilingual Language Processing
  • List of Corpora and Tools for Multilingual Language Processing
  • Example-based Machine Translation: A Survey
  • The Web As a Parallel Corpus: presentation of Resnik & Smith's Computational Linguistics '03 article
  • Machine Transliteration: review of Knight and Graehl's ACL 1997 paper
  • Improving SMT with Automatically Learned Rewrite Patterns: review of Xia and McCord's COLING 2004 paper
• Speech Recognition and Language Modeling
  • Factored Language Models Tutorial: lecture slides for EE517
  • BibTeX reading list for: Language Model Adaptation
  • The EM Algorithm and Discriminative Training with MMIE Criteria tutorial
  • Sound Perception and Models of Speech Production lecture notes
• Other
  • Gene and genome evolution slides (UW-CS molecular bio reading group): material based off Alberts' Essential Molecular Biology text.

Previous Projects

• Machine Learning for Operating Systems, at Microsoft Research
• Dependency Parsing with Bayes Point Machines at Microsoft Research
• Singing-voice synthesis at NTT Labs, Japan
• Multiscale image processing with Wedgelets at the Rice University DSP Group
• GPS bias estimation at the MIT Lincoln Laboratory
• A couple of my friends and I designed and fabricated a Fuzzy Logic processor chip, which won the AMD VLSI Design contest at Rice. See the chip photo and project website

Recent Activities

• I'm co-teaching EE511, Statistical Learning in Spring 2008 with Prof. Mari Ostendorf
• Organizer/Co-Chair, ACL/COLING 2006 Student Research Workshop
• Program Committee / Reviewer,
  • IEEE Transactions on Audio, Speech, and Language Processing
  • EMNLP 2008 (area: machine learning, machine translation)
  • ACL 2008 (area: machine translation)
  • ACM Transactions on Asian Language Information Processing
  • IEEE Potentials Magazine
  • EMNLP 2007 (areas: multilinguality, document collections)
  • NAACL Human Language Technology 2007 (area: machine translation)
  • Int'l Conf. on AI & Statistics (AISTATS) 2007
  • PACLING 2007
  • ACL/COLING 2006 (area: speech and language modeling)
  • IEEE Int'l Symposium on Signal Processing and Information Technology (2006)

Links

• My Personal Homepage
• Some interesting seminars: Machine Translation Reading Group , Graphical Models Reading Group , Markovia, NLP Reading Group.
• Download the Fortune Cookie Message Corpus, courtesy of Happy Panda Lucky Dragon Garden (c) and General Tsao's Chicken. Click here for our submission about this work to the Un-Natural Language Processing workshop held on April 1, 2005.

Back to the SSLI Lab homepage
Copyright ©1995-2008 University of Washington, College of Engineering - SSLI Lab
(Most Recent Update: Apr 15, 2008)