Paper: Fully Lexicalising CCGbank with Hat Categories

ACL ID D09-1126
Title Fully Lexicalising CCGbank with Hat Categories
Venue Conference on Empirical Methods in Natural Language Processing
Session Main Conference
Year 2009

We introduce an extension to CCG that al- lows form and function to be represented simultaneously, reducing the proliferation of modifier categories seen in standard CCG analyses. We can then remove the non-combinatory rules CCGbank uses to address this prob- lem, producingagrammarthatisfullylex- icalised and far less ambiguous. There are intrinsic benefits to full lexi- calisation, such as semantic transparency and simpler domain adaptation. The clear- est advantage is a 52-88% improvement in parse speeds, which comes with only a small reduction in accuracy.