Skip to main navigation Skip to search Skip to main content

Learning Domain Invariant Word Representations for Parsing Domain Adaptation

  • Xiuming Qiao*
  • , Yue Zhang
  • , Tiejun Zhao
  • *Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

We show that strong domain adaptation results for dependency parsing can be achieved using a conceptually simple method that learns domain-invariant word representations. Lacking labeled resources, dependency parsing for low-resource domains has been a challenging task. Existing work considers adapting a model trained on a resource-rich domain to low-resource domains. A mainstream solution is to find a set of shared features across domains. For neural network models, word embeddings are a fundamental set of initial features. However, little work has been done investigating this simple aspect. We propose to learn domain-invariant word representations by fine-tuning pretrained word representations adversarially. Our parser achieves error reductions of 5.6% UAS, 7.9% LAS on PTB respectively, and 4.2% UAS, 3.2% LAS on Genia respectively, showing the effectiveness of domain invariant word representations for alleviating lexical bias between source and target data.

Original languageEnglish
Title of host publicationNatural Language Processing and Chinese Computing - 8th CCF International Conference, NLPCC 2019, Proceedings
EditorsJie Tang, Min-Yen Kan, Dongyan Zhao, Sujian Li, Hongying Zan
PublisherSpringer
Pages801-813
Number of pages13
ISBN (Print)9783030322328
DOIs
StatePublished - 2019
Externally publishedYes
Event8th CCF International Conference on Natural Language Processing and Chinese Computing, NLPCC 2019 - Dunhuang, China
Duration: 9 Oct 201914 Oct 2019

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11838 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference8th CCF International Conference on Natural Language Processing and Chinese Computing, NLPCC 2019
Country/TerritoryChina
CityDunhuang
Period9/10/1914/10/19

Keywords

  • Dependency parsing
  • Domain adaptation
  • Generative Adversarial Network
  • Wasserstein distance
  • Word representations

Fingerprint

Dive into the research topics of 'Learning Domain Invariant Word Representations for Parsing Domain Adaptation'. Together they form a unique fingerprint.

Cite this