Abstract：Recent work on joint word segmentation, POS tagging, and dependency parsing in Chinese has two key problems: one is that the word segmentation based on character and the dependency parsing based on word are not well-combined in the transition-based framework; the other is that the current joint model suffers from the insufficiency of annotated corpus. In order to resolve the first problem, we propose to transform the conventional word-based dependency tree into character-based dependency tree by using the internal structure of words and then propose a novel character-level joint model for the three tasks. For Chinese word segmentation, we design 4 transition actions: Shfit_S, Shift_B, Shift_M and Shift_E, through which the features used in previous researches can also be integrated into the model. In order to resolve the second problem, we propose a novel semi-supervised joint model for exploiting n-gram feature and dependency subtree feature from partially-annotated corpus. Experimental results on the Chinese Treebank show that our joint model achieved the F1-scores of 98.31%, 94.84% and 81.71% for Chinese word segmentation, POS tagging, and dependency parsing, respectively. Our model outperforms the pipeline model in the three tasks by 0.92%, 1.77% and 3.95%, respectively. Especially, the F1 value of word segmentation and POS tagging achieved the best among the public results so far.
[1] Zhang Y, Clark S. A fast decoder for joint word segmentation and POS-tagging using a single discriminative model[C]//Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, 2010: 843-852. [2] Kruengkrai C, Uchimoto K, Kazama J, et al. An error-driven word-character hybrid model for joint Chinese word segmentation and POS tagging[C]//Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP: Volume 1-Volume 1. Association for Computational Linguistics, 2009: 513-521. [3] Hatori J, Matsuzaki T, Miyao Y, et al. Incremental Joint POS Tagging and Dependency Parsing in Chinese[C]//Proceedings of the IJCNLP. 2011: 1216-1224. [4] Li Z, Zhang M, Che W, et al. Joint models for Chinese POS tagging and dependency parsing[C]//Proceedings of the Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, 2011: 1180-1191. [5] Hatori J, Matsuzaki T, Miyao Y, et al. Incremental joint approach to word segmentation, pos tagging, and dependency parsing in chinese[C]//Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics: Long Papers-Volume 1. Association for Computational Linguistics, 2012: 1045-1053. [6] Zhang M, Zhang Y, Che W, et al. Chinese parsing exploiting characters[C]//Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics. 2013:125-134. [7] Guo Z, Zhang Y, Su C, et al. Exploration of N-gram Features for the Domain Adaptation of Chinese Word Segmentation[C]//Proceedings of the NLPCC 2012. [8] Wang Y, Jun'ichi Kazama Y T, Tsuruoka Y, et al. Improving Chinese Word Segmentation and POS Tagging with Semi-supervised Methods Using Large Auto-Analyzed Data[C]//Proceedings of the IJCNLP. 2011: 309-317. [9] Koo T, Carreras X, Collins M. Simple semi-supervised dependency parsing[C]//Proceedings of ACL/HLT. 2008:595 603. [10] Chen W, Kazama J, Uchimoto K, et al. Improving dependency parsing with subtrees from auto-parsed data[C]//Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing: Volume 2-Volume 2. Association for Computational Linguistics, 2009: 570-579. [11] Chen W, Kazama J, Torisawa K. Bitext dependency parsing with bilingual subtree constraints[C]//Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics. Association for Computational Linguistics, 2010: 21-29. [12] Zhang Y, Nivre J. Analyzing the Effect of Global Learning and Beam-Search on Transition-Based Dependency Parsing[C]//Proceedings of the COLING (Posters). 2012: 1391-1400. [13] Collins M, Roark B. Incremental parsing with the perceptron algorithm[C]//Proceedings of the 42nd Annual Meeting on Association for Computational Linguistics. Association for Computational Linguistics, 2004: 111. [14] Zhu M, Zhang Y, Chen W, et al. Fast and Accurate Shift-Reduce Constituent Parsing[C]//Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics. 2013: 434-443. [15] Li Z, Zhou G. Unified dependency parsing of Chinese morphological and syntactic structures[C]//Proceedings of the 2012 Joint Conference on Empirical Methods in Natural Language Processing and Computational Natural Language Learning. Association for Computational Linguistics, 2012: 1445-1454. [16] Zhao H, Huang C N, Li M, et al. Effective tag set selection in Chinese word segmentation via conditional random field modeling[C]//Proceedings of PACLIC. 2006, 20: 87-94. [17] McDonald R, Crammer K, Pereira F. Online large-margin training of dependency parsers[C]//Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics. Association for Computational Linguistics, 2005: 91-98. [18] Sun W. A stacked sub-word model for joint Chinese word segmentation and part-of-speech tagging[C]//Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies-Volume 1. Association for Computational Linguistics, 2011: 1385-1394. [19] Zhang M, Zhang Y, Che W, et al. Character-Level Chinese Dependency Parsing[C]//Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics. 2014: 1326 1336.
2014, Vol. 28 
