框架识别是语义角色标注的基本任务,它是根据目标词激起的语义场景,为其分配一个合适的语义框架。目前框架识别的研究主要是基于统计机器学习方法,把它看作多分类问题,框架识别的性能主要依赖于人工选择的特征。然而,人工选择特征的有效性和完备性无法保证。深度神经网络自动学习特征的能力,为我们提供了新思路。该文探索了利用深度神经网络自动学习目标词上下文特征,建立了一种新的通用的框架识别模型,在汉语框架网和《人民日报》2003年3月新闻语料上分别取得了79.64%和78.58%的准确率,实验证明该模型具有较好的泛化能力。
Abstract
Frame identification is a basic task of semantic role labeling, which assigns a correct frame to the labeled target word based on the semantic scene. At present, the state-of-the-art methods are primarily based on statistical machine learning, in which the performance heavily depends on the quality of the extracted features. This paper proposes a DNN based frame identification method, trying to capture the target word context automatically. Experiments on the Chinese FrameNet and the Peoples Daily(March, 2003) show 79.64% and 78.58% accuracy, respectively.
关键词
汉语框架 /
框架识别 /
深度神经网络 /
分布式表征
{{custom_keyword}} /
Key words
Chinese FramNet /
frame identification /
deep neural network /
distributed representation
/
/
/
/
/
/
/
/
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 李济洪.汉语框架语义角色的自动标注技术研究: [D].太原: 山西大学博士学位论文,2010.
[2] Ken Litkowski. CLR: Integration of FrameNet in a Text Representation System[C]//Proceedings of the 4th International Workshop on Semantic Evaluations. Prague, Czech Republic, 2007: 113-116.
[3] Ru Li, Haijing Liu, Shuanghong Li.Chinese Frame Identification using T-CRF Model[C]//Proceedings of International Conference on Computional Linguistics. Beijing, 2010: 674-682.
[4] Cosmin Adrian Bejan, Hathaway Chris. UTD-SRL: A pipeline Architecture for Extracting Frame Semantic Structures[C]//Proceedings of the 4th International Workshop on Semantic Evaluations.Prague, 2007: 460-463.
[5] C Baker, M Ellsworth, K Erk. SemEval-2007 Task 19: Frame Semantic Structure Extraction[C]//Proceedings of the 4th International Workshop on Semantic Evaluations.Prague, 2007: 99-104.
[6] Karl Moritz Hermann, Dipanjan Das, Jason Weston, et al. Semantic Frame Identification with Distributed Word Representations[C]//Proceedings of ACL 2014 Baltimore, USA. 2014: 1448-1458.
[7] 刘开瑛.汉语框架语义网(CFN)构建现状[C]//第四届 全国学生计算语言学研讨会会议论文集.2008: 1-7.
[8] C J Fillmore. Frame Semantics[J]. Linguistics in the Moring Calm, Hanshin Publishing Co.. Seoul, South Korea. 1982: 111-137.
[9] Burchardt A, Erk K, Frank A. A WordNet detour to FrameNet[C]//Proceedings of the GLDV 2005 Germa-Net II Workshop Bonn, Germany,2005.
[10] R Johansson, P Nugues.Using WordNet to extend FrameNet coverage[C]//Proceedings of the workshop on Building Frame-semantic Resources for Scandinavian and Baltic Languages.Tartu,2007.
[11] M Pennacchiotti, D De Cao, R Basili, et al.Automatic induction of FrameNet lexical units[C]//Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing. Honolulu,2008: 457-465.
[12] Dipanjan Das, Noah A Smith. Semi-Supervised Frame-Semantic Parsing for Unknown Predicate[C]//Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics. Portland, Oregon, 2011: 1435-1444.
[13] 陈雪丽,李茹,王赛等.汉语框架网中未登录词元的框架选择[J].中文信息学报.2014,28(3): 48-54,61.
[14] Cosmin Adrian Bejan, Hathaway Chris .UTD-SRL: A Pipeline Architecture for Extracting Frame Semantic Structures[C]//Proceedings of 45th annual meeting of Association for Computational Linguistics, 2007: 460-463.
[15] Richard Johansson, Nugues Pierre.LTH: Semantic Structure Extraction using Nonprojective Dependency Trees[C]//Proceedings of the 4th International Work on Semantic Evaluations. Prague, 2007: 227-230.
[16] 李国臣,张立凡,李茹等.基于词元语义特征的汉语框架排歧研究[J].中文文信息学报.2013,27(4): 44-51.
[17] 哈尔滨工业大学LTP平台[CP]. http://www.ltp-cloud.com/document/#api_rest_note
[18] 党帅兵,李国臣,王瑞波等. 基于词分布表征的汉语框架排歧研究[J].中北大学学报.2015,36(3): 328-332,337.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
基金
国家自然科学基金(61373082);国家863计划(2015AA015407);山西省科技基础条件平台建设项目(2014091004-0103);山西省回国留学人员科研资助项目(2013-015);中国民航大学信息安全测评中心开放课题基金(CAAC-ISECCA-201402);国家自然科学基金(61673248)
{{custom_fund}}
PDF(3700 KB)
