Density-Driven Cross-Lingual Transfer of Dependency Parsers

Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 328–338, Lisbon, Portugal, 17-21 September 2015. c⃝2015 Association for Computational Linguistics. Density-Driven Cross-Lingual Transfer of Dependency Parsers Mohammad Sadegh Rasooli and Michael Collins∗ Department of Computer Science, Columbia University New York, NY 10027, USA {rasooli,mcollins}@cs.columbia.edu Abstract We present a novel method for the cross- lingual transfer of dependency parsers. Our goal is to induce a dependency parser in a target language of interest without any direct supervision: instead we as- sume access to parallel translations be- tween the target and one or more source languages, and to supervised parsers in the source language(s). Our key contribu- tions are to show the utility of dense pro- jected structures when training the target language parser, and to introduce a novel learning algorithm that makes use of dense structures. Results on several languages show an absolute improvement of 5.51% in average dependency accuracy over the state-of-the-art method of (Ma and Xia, 2014). Our average dependency accuracy of 82.18% compares favourably to the ac- curacy of fully supervised methods. 1 Introduction In recent years there has been a great deal of inter- est in dependency parsing models for natural lan- guages. Supervised learning methods have been shown to produce highly accurate dependency- parsing models; unfortunately, these methods rely on human-annotated data, which is expensive to obtain, leading to a significant barrier to the devel- opment of dependency parsers for new languages. Recent work has considered unsupervised meth- ods (e.g. (Klein and Manning, 2004; Headden III et al., 2009; Gillenwater et al., 2011; Mareˇcek and Straka, 2013; Spitkovsky et al., 2013; Le and Zuidema, 2015; Grave and Elhadad, 2015)), or methods that transfer linguistic structures across languages (e.g. (Cohen et al., 2011; McDonald et al., 2011; Ma and Xia, 2014; Tiedemann, 2015; ∗Currently on leave at Google Inc. New York. Guo et al., 2015; Zhang and Barzilay, 2015; Xiao and Guo, 2015)), in an effort to reduce or eliminate the need for annotated training examples. Unfor- tunately the accuracy of these methods generally lags quite substantially behind the performance of fully supervised approaches. This paper describes novel methods for the transfer of syntactic information between lan- guages. As in previous work (Hwa et al., 2005; Ganchev et al., 2009; McDonald et al., 2011; Ma and Xia, 2014), our goal is to induce a dependency parser in a target language of interest without any direct supervision (i.e., a treebank) in the target language: instead we assume access to parallel translations between the target and one or more source languages, and to supervised parsers in the source languages. We can then use alignments in- duced using tools such as GIZA++ (Och and Ney, 2000), to transfer dependencies from the source language(s) to the target language (example pro- jections are shown in Figure 1). A target language parser is then trained on the projected dependen- cies. Our contributions are as follows: • We demonstrate the utility of dense projected structures when training the target-language parser. …

Content truncated to 500 words