Understanding the Effect of Textual Adversaries in Multimodal Machine Translation

Koel Dutta Chowdhury, Desmond Elliott


Abstract
It is assumed that multimodal machine translation systems are better than text-only systems at translating phrases that have a direct correspondence in the image. This assumption has been challenged in experiments demonstrating that state-of-the-art multimodal systems perform equally well in the presence of randomly selected images, but, more recently, it has been shown that masking entities from the source language sentence during training can help to overcome this problem. In this paper, we conduct experiments with both visual and textual adversaries in order to understand the role of incorrect textual inputs to such systems. Our results show that when the source language sentence contains mistakes, multimodal translation systems do not leverage the additional visual signal to produce the correct translation. We also find that the degradation of translation performance caused by textual adversaries is significantly higher than by visual adversaries.
Anthology ID:
D19-6406
Volume:
Proceedings of the Beyond Vision and LANguage: inTEgrating Real-world kNowledge (LANTERN)
Month:
November
Year:
2019
Address:
Hong Kong, China
Venues:
EMNLP | WS
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
35–40
URL:
https://www.aclweb.org/anthology/D19-6406
DOI:
10.18653/v1/D19-6406
Bib Export formats:
BibTeX MODS XML EndNote
PDF:
https://www.aclweb.org/anthology/D19-6406.pdf