Building a treebank for Vietnamese dependency parsing

Abstract

The problem of Vietnamese syntactic parsing, especially constituency parsing, has recently been tackled by several research groups. A common effort of the Vietnamese language processing community has allowed the creation of VietTreebank, a reference parsed corpus containing about 10,000 sentences for the constituency parsing task. In this paper, we present our work to build a reference treebank, based on VietTreebank, for the dependency parsing task, which has not yet been very well studied for Vietnamese. First we define a dependency label set by adapting the dependency schema developed by the NLP group at Stanford university and taking into account the particularities of Vietnamese grammar. Then we propose an algorithm to convert a constituency treebank to a dependency one. The algorithm is tested on a set of 100 sentences of VietTreebank corpus and gives very good results. Finally, we carry out an experiment on Vietnamese dependency parsing using MaltParser tool and the dependency treebank converted from VietTreebank.

Error Analysis for Vietnamese Dependency Parsing

“…There are only a few studies on dependency parsing for the Vietnamese language. The works [2], [8] built a dependency treebank for the Vietnamese language by converting the Vietnamese Treebank from phrase structures to dependency structures. The highest accuracies of MSTParser and MaltParser trained on this treebank are 71.66% (LAS) and 70.49% (LAS) respectively [2].…”

“…State-of-the-art methods on dependency parsing for the Vietnamese language achieved only less than 80% [2], [8]. To understand why parsing performance is low and know how to improve it, we carried out a thorough error analysis based on two state-of-the-art data-driven dependency parsing systems, MSTParser and MaltParser.…”

A syntactic component for Vietnamese language processing

“…However, the constituency parser and dependency parser developed in this work are currently purely symbolic in that they do not make use of any probabilistic evidence to discriminate good parses from bad ones for a given sentence, regardless of its grammaticality. An initial investigation of statistical dependency parsing for Vietnamese has shown encouraging results (Nguyen et al 2013). We believe that there is room to improve the performance of dependency parsers in general and of our dependency parser in particular by employing a hybrid approach: use elementary trees of an lexicalized tree-adjoining grammar as good syntactic features in a statistical dependency parser.…”

A Case Study on Meaning Representation for Vietnamese

“…• Vietnamese dependency treebank: in (Thi et al, 2013), the authors define a dependency label set based on the English dependency schema. Next, they propose an algorithm to transform more than 10000 sentences from VietTreebank into a dependency treebank (Phuong et al, 2015), (T-L et al, 2016).…”