We describe a transfer method based on annotation projection to develop a dependency-based semantic role labeling system for languages for which no supervised linguistic information other than parallel data is available. Unlike previous work that presumes the availability of supervised features such as lemmas, part-of-speech tags, and dependency parse trees, we only make use of word and character features. Our deep model considers using character-based representations as well as unsupervised stem embeddings to alleviate the need for supervised features. Our experiments outperform a state-of-the-art method that uses supervised lexico-syntactic features on 6 out of 7 languages in the Universal Proposition Bank.

No abstract

We present a novel method for the crosslingual transfer of dependency parsers. Our goal is to induce a dependency parser in a target language of interest without any direct supervision: instead we assume access to parallel translations between the target and one or more source languages, and to supervised parsers in the source language(s). Our key contributions are to show the utility of dense projected 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 accuracy of fully supervised methods.

Ezafe construction is an idiosyncratic phenomenon in the Persian language. It is a good indicator for phrase boundaries and dependency relations but mostly does not appear in the text. In this paper, we show that adding information about Ezafe construction can give 4.6% relative improvement in dependency parsing and 9% relative improvement in shallow parsing. For evaluation purposes, Ezafe tags are manually annotated in the Persian dependency treebank. Furthermore, to be able to conduct experiments on shallow parsing, we develop a dependency to shallow phrase structure convertor based on the Persian dependencies.

We present a novel way of generating unseen words, which is useful for certain applications such as automatic speech recognition or optical character recognition in low-resource languages. We test our vocabulary generator on seven low-resource languages by measuring the decrease in out-of-vocabulary word rate on a held-out test set. The languages we study have very different morphological properties; we show how our results differ depending on the morphological complexity of the language. In our best result (on Assamese), our approach can predict 29% of the token-based out-of-vocabulary with a small amount of unlabeled training data.