In this paper, we study domain adaptation of semantic role classification. Most systems utilize the supervised method for semantic role classification. But, these methods often suffer severe performance drops on out-of-domain test data. The reason for the performance drops is that there are giant feature differences between source and target domain. This paper proposes a framework called Adversarial Domain Adaption Network (ADAN) to relieve domain adaption of semantic role classification. The idea behind our method is that the proposed framework can derive domain-invariant features via adversarial learning and narrow down the gap between source and target feature space. To evaluate our method, we conduct experiments on English portion in the CoNLL 2009 shared task. Experimental results show that our method can largely reduce the performance drop on out-of-domain test data.
Haitong YANG
China Central Normal University
Guangyou ZHOU
China Central Normal University
Tingting HE
China Central Normal University
Maoxi LI
Jiangxi Normal University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Haitong YANG, Guangyou ZHOU, Tingting HE, Maoxi LI, "Adversarial Domain Adaptation Network for Semantic Role Classification" in IEICE TRANSACTIONS on Information,
vol. E102-D, no. 12, pp. 2587-2594, December 2019, doi: 10.1587/transinf.2019EDP7087.
Abstract: In this paper, we study domain adaptation of semantic role classification. Most systems utilize the supervised method for semantic role classification. But, these methods often suffer severe performance drops on out-of-domain test data. The reason for the performance drops is that there are giant feature differences between source and target domain. This paper proposes a framework called Adversarial Domain Adaption Network (ADAN) to relieve domain adaption of semantic role classification. The idea behind our method is that the proposed framework can derive domain-invariant features via adversarial learning and narrow down the gap between source and target feature space. To evaluate our method, we conduct experiments on English portion in the CoNLL 2009 shared task. Experimental results show that our method can largely reduce the performance drop on out-of-domain test data.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2019EDP7087/_p
Copy
@ARTICLE{e102-d_12_2587,
author={Haitong YANG, Guangyou ZHOU, Tingting HE, Maoxi LI, },
journal={IEICE TRANSACTIONS on Information},
title={Adversarial Domain Adaptation Network for Semantic Role Classification},
year={2019},
volume={E102-D},
number={12},
pages={2587-2594},
abstract={In this paper, we study domain adaptation of semantic role classification. Most systems utilize the supervised method for semantic role classification. But, these methods often suffer severe performance drops on out-of-domain test data. The reason for the performance drops is that there are giant feature differences between source and target domain. This paper proposes a framework called Adversarial Domain Adaption Network (ADAN) to relieve domain adaption of semantic role classification. The idea behind our method is that the proposed framework can derive domain-invariant features via adversarial learning and narrow down the gap between source and target feature space. To evaluate our method, we conduct experiments on English portion in the CoNLL 2009 shared task. Experimental results show that our method can largely reduce the performance drop on out-of-domain test data.},
keywords={},
doi={10.1587/transinf.2019EDP7087},
ISSN={1745-1361},
month={December},}
Copy
TY - JOUR
TI - Adversarial Domain Adaptation Network for Semantic Role Classification
T2 - IEICE TRANSACTIONS on Information
SP - 2587
EP - 2594
AU - Haitong YANG
AU - Guangyou ZHOU
AU - Tingting HE
AU - Maoxi LI
PY - 2019
DO - 10.1587/transinf.2019EDP7087
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E102-D
IS - 12
JA - IEICE TRANSACTIONS on Information
Y1 - December 2019
AB - In this paper, we study domain adaptation of semantic role classification. Most systems utilize the supervised method for semantic role classification. But, these methods often suffer severe performance drops on out-of-domain test data. The reason for the performance drops is that there are giant feature differences between source and target domain. This paper proposes a framework called Adversarial Domain Adaption Network (ADAN) to relieve domain adaption of semantic role classification. The idea behind our method is that the proposed framework can derive domain-invariant features via adversarial learning and narrow down the gap between source and target feature space. To evaluate our method, we conduct experiments on English portion in the CoNLL 2009 shared task. Experimental results show that our method can largely reduce the performance drop on out-of-domain test data.
ER -