Paper: Compiling Comp Ling: Weighted Dynamic Programming And The Dyna Language

ACL ID H05-1036
Title Compiling Comp Ling: Weighted Dynamic Programming And The Dyna Language
Venue Conference on Empirical Methods in Natural Language Processing
Session Main Conference
Year 2005
Authors

Weighted deduction with aggregation is a powerful theoretical formalism that encompasses many NLP algorithms. This pa- per proposes a declarative specification language, Dyna; gives general agenda-based algorithms for computing weights and gradients; briefly discusses Dyna-to-Dyna program transforma- tions; and shows that a first implementation of a Dyna-to-C++ compiler produces code that is efficient enough for real NLP re- search, though still several times slower than hand-crafted code.