The Unreasonable Volatility of Neural Machine Translation Models

Marzieh Fadaee, Christof Monz


Abstract
Recent works have shown that Neural Machine Translation (NMT) models achieve impressive performance, however, questions about understanding the behavior of these models remain unanswered. We investigate the unexpected volatility of NMT models where the input is semantically and syntactically correct. We discover that with trivial modifications of source sentences, we can identify cases where unexpected changes happen in the translation and in the worst case lead to mistranslations. This volatile behavior of translating extremely similar sentences in surprisingly different ways highlights the underlying generalization problem of current NMT models. We find that both RNN and Transformer models display volatile behavior in 26% and 19% of sentence variations, respectively.
Anthology ID:
2020.ngt-1.10
Volume:
Proceedings of the Fourth Workshop on Neural Generation and Translation
Month:
July
Year:
2020
Address:
Online
Editors:
Alexandra Birch, Andrew Finch, Hiroaki Hayashi, Kenneth Heafield, Marcin Junczys-Dowmunt, Ioannis Konstas, Xian Li, Graham Neubig, Yusuke Oda
Venue:
NGT
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
88–96
Language:
URL:
https://aclanthology.org/2020.ngt-1.10
DOI:
10.18653/v1/2020.ngt-1.10
Bibkey:
Cite (ACL):
Marzieh Fadaee and Christof Monz. 2020. The Unreasonable Volatility of Neural Machine Translation Models. In Proceedings of the Fourth Workshop on Neural Generation and Translation, pages 88–96, Online. Association for Computational Linguistics.
Cite (Informal):
The Unreasonable Volatility of Neural Machine Translation Models (Fadaee & Monz, NGT 2020)
Copy Citation:
PDF:
https://aclanthology.org/2020.ngt-1.10.pdf
Video:
 http://slideslive.com/38929823
Code
 marziehf/variation-generation