Neural Cross-Lingual Event Detection with Minimal Parallel Resources

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, pages 738–748, Hong Kong, China, November 3–7, 2019. c⃝2019 Association for Computational Linguistics 738 Neural Cross-Lingual Event Detection with Minimal Parallel Resources Jian Liu12∗, Yubo Chen1∗, Kang Liu12, Jun Zhao12 1 National Laboratory of Pattern Recognition, Institute of Automation Chinese Academy of Sciences, Beijing, 100190, China 2 University of Chinese Academy of Sciences {jian.liu, yubo.chen, kliu, jzhao}@nlpr.ia.ac.cn Abstract The scarcity in annotated data poses a great challenge for event detection (ED). Cross- lingual ED aims to tackle this challenge by transferring knowledge between different lan- guages to boost performance. However, pre- vious cross-lingual methods for ED demon- strated a heavy dependency on parallel re- sources, which might limit their applicability. In this paper, we propose a new method for cross-lingual ED, demonstrating a minimal de- pendency on parallel resources. Specifically, to construct a lexical mapping between differ- ent languages, we devise a context-dependent translation method; to treat the word order dif- ference problem, we propose a shared syntac- tic order event detector for multilingual co- training. The efficiency of our method is studied through extensive experiments on two standard datasets. Empirical results indicate that our method is effective in 1) perform- ing cross-lingual transfer concerning differ- ent directions and 2) tackling the extremely annotation-poor scenario. 1 Introduction Event detection (ED) is a crucial natural language processing problem that aims to identify event triggers in texts (Ahn, 2006; Nguyen and Grish- man, 2015). For example, in the sentence: “A man died when a tank fired on the hotel”, ED requires a system to identify two event triggers, died and fired, along with their types Die and Attack. Generally, training an ED system requires to obtain a considerably large amount of labeled data. However, owing to the complexity and high costs of annotation, existing event resources are scarce and unbalanced across languages (Hsi et al., 2016), which prevents us from building an ED system in languages with insufficient train- ing data. Cross-lingual ED (Ji, 2009; Chen and ∗Equal contribution. Ji, 2009; Zhu et al., 2014; Hsi et al., 2016; Liu et al., 2018a) aims to tackle this challenge by transferring knowledge cross languages to boost performance. However, previous cross-lingual ED methods rely on either high-performance machine translation (MT) systems trained on large num- bers of parallel sentences or manually aligned doc- uments to achieve a decent performance — the required parallel resources may only exist for a small fraction of language pairs (Koehn et al., 2007), which greatly limits the applicability of these methods. In this paper, we propose a new simple but ef- fective method for cross-lingual ED, which can overcome the data scarcity problem in annotation- poor languages by jointly training with resources in other languages. Compared with previous methods, our approach demonstrates a minimal dependency on parallel resources, which may fit with language pairs that do not have large bitexts. To achieve cross-lingual transfer, two chal- lenges …

Content truncated to 500 words