Yoav Artzi, Kenton Lee and Luke Zettlemoyer of UW CSE’s natural language processing group have captured a Best Paper Award at EMNLP 2015 – one of only two best papers selected from more than 600 submissions to the conference on empirical methods in natural language processing.
The paper, Broad-coverage CCG Semantic Parsing with AMR, describes an approach for learning parsers that build Abstract Meaning Representations (AMRs), a recently proposed, general formalism for representing core aspects of sentence meaning. The team’s work is a significant step over previous CCG grammar induction algorithms, as it learns to analyze a much wider range of sentence types and scales to much larger data sets. The paper will be presented at EMNLP 2015 in Lisbon, Portugal later this month.
This latest award continues UW CSE’s enviable winning streak at leading conferences, and is the second paper award for the NLP group in 2015 (just one shy of repeating the systems group’s hat trick). It is also a great start for Yoav’s new faculty career at Cornell!
Read the paper here.
Congratulations Yoav, Kenton and Luke!