Christian MonsonPost-Doctoral Fellow Center for Spoken Language Understanding
I am broadly interested in computational treatment of natural language — an area that includes everything from machine translation, natural language understanding, and parsing to information retrieval and speech recognition. I am currently working with Brian Roark and Zak Shafran on two natural language problems. First, I am extending my thesis work on unsupervised morphology induction. And Second, I am expanding an existing system for spoken term detection.
In the Spring of 2010 I organized NW-NLP, a Pacific Northwest Regional Speech and Natural Language Processing Workshop held at Microsoft Research. The 1-day workshop was a resounding success, with approximately 100 people in attendance from a half dozen institutions representing academia, industry, and government.
In December 2008 I graduated from the Language Technologies Institute at Carnegie Mellon University. My Ph.D. thesis describes ParaMor, an algorithm that automatically discovers the morphological structure of a language from nothing more than raw text from that language. Morpho Challenge is a peer-operated competition for unsupervised morphology induction algorithms. In the 2009 Morpho Challenge competition, the ParaMor algorithm placed first at F1 of morpheme identification in four of six language tracks. My thesis advisors at Carnegie Mellon were Jaime Carbonell, Alon Lavie , and Lori Levin, With Ron Kaplan of PowerSet rounding out my thesis committee.