Improving Feature-Rich Transition-Based Constituent Parsing Using Recurrent Neural Networks

Improving Feature-Rich Transition-Based Constituent Parsing Using Recurrent Neural Networks
复制标题

使用循环神经网络改进基于特征丰富的转换的成分解析

DOI:
10.1587/transinf.2017edp7003
复制
发表时间:
2017-09
影响因子:
0.7
通讯作者:
Sumita Eiichiro
Sumita Eiichiro
中科院分区:
计算机科学4区
文献类型:
--
作者:
Ma Chunpeng;Tamura Akihiro;Liu Lemao;Zhao Tiejun;Sumita Eiichiro

文献摘要

参考文献

相似文献

传统的基于手动调优特性的功能丰富的解析器已经达到了最先进的性能。然而,这些解析器并不擅长仅使用准备好的特性模板捕获的线索来处理长期依赖关系。另一方面,基于递归神经网络(RNN)的解析器可以有效地编码无界的历史信息,但对于小的树状结构,特别是涉及低频词时,它们的表现不佳,而且它们不能使用先验的语言知识。在本文中,我们提出了一个简单而有效的框架来结合特征丰富的基于转换的解析器和rnn的优点。具体来说,所提出的框架将基于rnn的分数整合到特征丰富的解析器使用的特征模板中。在英文WSJ树库和SPMRL 2014德语树库上,我们的框架达到了最先进的性能(英语为91.56 f分,德语为83.06 f分),而不需要任何额外的未标记数据。关键词:成分解析,递归神经网络,系统组合
Conventional feature-rich parsers based on manually tuned features have achieved state-of-the-art performance. However, these parsers are not good at handling long-term dependencies using only the clues captured by a prepared feature template. On the other hand, recurrent neural network (RNN)-based parsers can encode unbounded history information effectively, but they perform not well for small tree structures, especially when low-frequency words are involved, and they cannot use prior linguistic knowledge. In this paper, we propose a simple but effective framework to combine the merits of feature-rich transition-based parsers and RNNs. Specifically, the proposed framework incorporates RNN-based scores into the feature template used by a feature-rich parser. On English WSJ treebank and SPMRL 2014 German treebank, our framework achieves state-of-the-art performance (91.56 F-score for English and 83.06 F-score for German), without requiring any additional unlabeled data. key words: constituent parsing, recurrent neural network, system combination
DOI: 10.18653/v1/p16-2093
发表时间: 2016-08
影响因子: 7
作者:
Joël Legrand;R. Collobert
通讯作者: Joël Legrand;R. Collobert
DOI: 10.18653/v1/d16-1001
发表时间: 2016-11
期刊: ArXiv
影响因子: --
作者:
James Cross;Liang Huang
通讯作者: James Cross;Liang Huang
DOI: --
发表时间: 2013-10
期刊: --
影响因子: --
作者:
Anders Björkelund;Özlem Çetinoğlu;Richárd Farkas;Thomas Müller;Wolfgang Seeker
通讯作者: Anders Björkelund;Özlem Çetinoğlu;Richárd Farkas;Thomas Müller;Wolfgang Seeker
DOI: 10.3115/1073445.1073459
发表时间: 2003-05
期刊: Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology - NAACL '03
影响因子: --
作者:
James Henderson
通讯作者: James Henderson
DOI: 10.18653/v1/d16-1257
发表时间: 2016-11
期刊: --
影响因子: --
作者:
Do Kook Choe;Eugene Charniak
通讯作者: Do Kook Choe;Eugene Charniak