A Polynomial-Time Dynamic Oracle for Non-Projective Dependency Parsing
Carlos Gómez‐Rodríguez, Francesco Sartorio, Giorgio Satta · 2014
The introduction of dynamic oracles has considerably improved the accuracy of greedy transition-based dependency parsers, without sacrificing parsing efficiency.However, this enhancement is limited to projective parsing, and dynamic oracles have not yet been implemented for parsers supporting non-projectivity.In this paper we introduce the first such oracle, for a non-projective parser based on Attardi's parser.We show that training with this oracle improves parsing accuracy over a conventional (static) oracle on a wide range of datasets.