Multilingual Transformer Encoders: a Word-Level Task-Agnostic Evaluation
Gaschi, Félix, Plesse, François, Rastin, Parisa, Toussaint, Yannick
–arXiv.org Artificial Intelligence
Some Transformer-based models can perform cross-lingual transfer learning: those models can be trained on a specific task in one language and give relatively good results on the same task in another language, despite having been pre-trained on monolingual tasks only. But, there is no consensus yet on whether those transformer-based models learn universal patterns across languages. We propose a word-level task-agnostic method to evaluate the alignment of contextualized representations built by such models. We show that our method provides more accurate translated word pairs than previous methods to evaluate word-level alignment. And our results show that some inner layers of multilingual Transformer-based models outperform other explicitly aligned representations, and even more so according to a stricter definition of multilingual alignment.
arXiv.org Artificial Intelligence
Jul-19-2022
- Country:
- Asia > China
- Hong Kong (0.04)
- Europe
- Finland > Southwest Finland
- Turku (0.04)
- France
- Grand Est > Meurthe-et-Moselle
- Nancy (0.04)
- Île-de-France > Paris
- Paris (0.04)
- Grand Est > Meurthe-et-Moselle
- Netherlands > South Holland
- The Hague (0.04)
- Sweden > Östergötland County
- Linköping (0.04)
- Finland > Southwest Finland
- North America
- Canada (0.04)
- United States > Oregon
- Multnomah County > Lents (0.04)
- Asia > China
- Genre:
- Research Report > New Finding (0.69)
- Technology: