David Burkett



I used to be a graduate student and then a postdoctoral researcher in Computer Science at the University of California, Berkeley, working in the Natural Language Processing Group with Professor Dan Klein. After leaving Berkeley, I worked on Search Quality at Twitter for a couple of years and am now at Semantic Machines.
Long ago, I also did some work in game AI and in this area am probably best known for my work on the Overmind, an agent for playing StarCraft.
Contact
dburkett ATSIGN cs DOT berkeley DOT edu
Research
I am interested in all sorts of things involving language, computers, and statistics, but my main interest is in how different languages relate to each other, both historically and linguistically. So far I've mostly been focused on work related to syntactic relationships between different languages. This means I've done some syntactic machine translation, but the majority of my work has been related to bilingual parsing.
Teaching
 In Fall 2009, I was a GSI for CS188: Introduction to Artificial Intelligence.
 In Spring 2010, I helped in the design and management of CS19413: LargeScale Decision Making, a new course covering advanced AI topics and including all the students in a course project to build a Starcraft agent. This agent became a prototype for our competition bot.
 In Spring 2011, I was a GSI for the revamped CS19413, where we reworked the class to have a more formal lecture and project structure. Students completed projects in advanced AI topics and integrated those lessons into the construction of their own Starcraft agents.
 I was a recipient of Berkeley's Outstanding GSI Award in 20092010, and the Computer Science Division's Outstanding GSI Award in 2012.
 I gave a tutorial with Dan Klein called Variational Inference in Structured NLP Models at NAACL 2012 and ACL 2013.
Publications
 David Hall, Alon Cohen, David Burkett, and Dan Klein, "Faster Optimal Planning with PartialOrder Pruning". Proceedings of ICAPS, 2013. [pdf]
 David Burkett and Dan Klein, "Transforming Trees to Improve Syntactic Convergence". Proceedings of EMNLP, 2012. [pdf]
[slides (key) (pdf)]
 Taylor BergKirkpatrick, David Burkett, and Dan Klein, "An Empirical Investigation of Statistical Significance in NLP". Proceedings of EMNLP, 2012.
 David Burkett and Dan Klein, "Fast Inference in Phrase Extraction Models with Belief Propagation". Proceedings of NAACL, 2012.
[pdf]
[slides (key) (pdf)]
 David Burkett, David Hall, and Dan Klein. "Optimal Graph Search with Iterated Graph Cuts". Proceedings of AAAI, 2011.
[pdf]
 Jonathan K. Kummerfeld, Mohit Bansal, David Burkett, and Dan Klein. "Mention Detection: Heuristics for the OntoNotes Annotations". Proceedings of CoNLL: Shared Task, 2011.
[pdf]
[poster]
 David Burkett, Slav Petrov, John Blitzer, and Dan Klein. "Learning Better Monolingual Models with Unannotated Bilingual Text". Proceedings of CoNLL, 2010.
[pdf]
[slides (key) (pdf) (mov)]
 David Burkett, John Blitzer, and Dan Klein. "Joint Parsing and Alignment with Weakly Synchronized Grammars". Proceedings of NAACL, 2010.
[pdf]
[slides (key) (pdf)]
 David Burkett and Tom Griffiths. "Iterated Learning of Multiple Languaged from Multiple Teachers". Proceedings of Evolang, 2010.
[pdf]
[slides (key) (pdf)]
 Jing Zheng, Necip Fazil Ayan, Wen Wang, and David Burkett. "Using Syntax in Largescale Audio Document Translation". Proceedings of Interspeech, 2009.
[pdf]
 David Burkett and Dan Klein. "Two Languages are Better than one (for Syntactic Parsing)". Proceedings of EMNLP, 2008.
[pdf]
[slides (key) (pdf)]