Ph.D. 2016, University of California, Berkeley

Natural Language Processing
Machine Learning

Greg Durrett will be joining UT Computer Science as an Assistant Professor in Fall 2017.

Greg's research focuses on solving core natural language processing problems, all of which are fundamentally concerned with turning unstructured text into structured information.  This kind of text processing is a critical step for allowing computers to access all of the information that's available on the web.  To tackle this problem, Greg's work uses structured machine learning methods, especially joint models that integrate multiple approaches or address multiple tasks simultaneously.  Such models need to be sophisticated and high-capacity so they can make use of large datasets, yet also tailored to capture the key linguistic phenomena specific to each task.  Greg has studied a range of NLP problems including coreference resolution, entity linking, syntactic parsing, and document summarization.

Greg values building systems that can be deployed outside the laboratory.  He has previously worked on conversational dialogue systems at Semantic Machines and on machine translation at Google.  He also maintains several publicly available NLP tools that are byproducts of his research.

Prior to joining UT Austin, Greg completed his Ph.D. at UC Berkeley, where he was advised by Dan Klein.  During his graduate career, Greg received the Facebook Fellowship in Natural Language Processing and an NSF Graduate Research Fellowship.  His work on coreference resolution was a Best Paper finalist at EMNLP 2013.