Adapting to Non-Centered Languages for Zero-shot Multilingual Translation
–arXiv.org Artificial Intelligence
Multilingual neural machine translation can translate unseen language pairs during training, i.e. zero-shot translation. However, the zero-shot translation is always unstable. Although prior works attributed the instability to the domination of central language, e.g. English, we supplement this viewpoint with the strict dependence of non-centered languages. In this work, we propose a simple, lightweight yet effective language-specific modeling method by adapting to non-centered languages and combining the shared information and the language-specific information to counteract the instability of zero-shot translation. Experiments with Transformer on IWSLT17, Europarl, TED talks, and OPUS-100 datasets show that our method not only performs better than strong baselines in centered data conditions but also can easily fit non-centered data conditions. By further investigating the layer attribution, we show that our proposed method can disentangle the coupled representation in the correct direction.
arXiv.org Artificial Intelligence
Sep-9-2022
- Country:
- Africa > Niger (0.04)
- Asia
- China > Hong Kong (0.04)
- Middle East > Saudi Arabia
- Ḥaʼil Province > Ha'il (0.04)
- Europe
- North America > United States
- Louisiana > Orleans Parish
- New Orleans (0.04)
- Minnesota > Hennepin County
- Minneapolis (0.14)
- New Mexico > Santa Fe County
- Santa Fe (0.04)
- Pennsylvania (0.04)
- Louisiana > Orleans Parish
- Oceania > Australia
- Genre:
- Research Report (0.40)
- Technology: