Specializing Multi-domain NMT via Penalizing Low Mutual Information
Jiyoung Lee, Hantae Kim, Hyunchang Cho, Edward Choi, Cheonbok Park
Abstract
Multi-domain Neural Machine Translation (NMT) trains a single model with multiple domains. It is appealing because of its efficacy in handling multiple domains within one model. An ideal multi-domain NMT learns distinctive domain characteristics simultaneously, however, grasping the domain peculiarity is a non-trivial task. In this paper, we investigate domain-specific information through the lens of mutual information (MI) and propose a new objective that penalizes low MI to become higher.Our method achieved the state-of-the-art performance among the current competitive multi-domain NMT models. Also, we show our objective promotes low MI to be higher resulting in domain-specialized multi-domain NMT.
BibTeX
@inproceedings{lee-etal-2022-specializing,
title = "Specializing Multi-domain {NMT} via Penalizing Low Mutual Information",
author = "Lee, Jiyoung and
Kim, Hantae and
Cho, Hyunchang and
Choi, Edward and
Park, Cheonbok",
editor = "Goldberg, Yoav and
Kozareva, Zornitsa and
Zhang, Yue",
booktitle = "Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing",
month = dec,
year = "2022",
address = "Abu Dhabi, United Arab Emirates",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2022.emnlp-main.680/",
doi = "10.18653/v1/2022.emnlp-main.680",
pages = "10015--10026"
}