Output details
11 - Computer Science and Informatics
University of Sheffield
A Gibbs Sampler for Phrasal Synchronous Grammar Induction
<24>The paper was pioneering work, framing automatic machine translation (MT) as grammar induction, with demonstrated improvements over state-of-the-art. This paper launched Cohn’s work on bilingual grammar and finite state transducer induction with subsequent papers at NIPS09, EMNLP09, NAACL10, 2xACL13. The paper has 62 citations (Google Scholar) and has been instrumental in the widespread adoption of Bayesian non-parametric techniques in MT and language processing (formed basis of tutorial by world foremost MT researcher, Kevin Knight). This work resulted in funding to run a JHU summer research workshop 2010 (NSF, Microsoft, Google sponsored), and for MT projects (EPSRC EP/I034750/1, Google project SLATR).