Performance-oriented dependency parsing

Abstract

In the last decade a lot of dependency parsers have been developed. This book describes the motivation for the development of yet another parser - MDParser. The state of the art is presented and the deficits of the current developments are discussed. The main problem of the current parsers is that the task of dependency parsing is treated independently of what happens before and after it. However, in practice parsing is rarely done for the sake of parsing itself, but rather in order to use the results in a follow-up application. Additionally, current parsers are accuracy-oriented and focus only on the quality of the results, neglecting other important properties, especially efficiency. The evaluation of some NLP technologies is sometimes as difficult as the task itself. For dependency parsing it was long thought not to be the case, however, some recent works show that the current evaluation possibilities are limited. This book proposes a methodology to account for the weaknesses and combine the strengths of the current approaches. Finally, MDParser is evaluated against other state-of-the-art parsers. The results show that it is the fastest parser currently available and it is able to process plain text, which other parsers usually cannot. The results are slightly behind the top accuracies in the field, however, it is demonstrated that it is not decisive for applications.