Machine Translation Evaluation with BERT Regressor

Abstract

We introduce the metric using BERT (Bidirectional Encoder Representations from Transformers) (Devlin et al., 2019) for automatic machine translation evaluation. The experimental results of the WMT-2017 Metrics Shared Task dataset show that our metric achieves state-of-the-art performance in segment-level metrics task for all to-English language pairs.

arXiv comment6 pages

References in corpus (1)

• An efficient framework for learning sentence representations

Cited by in corpus (5)

• Unbabel's Participation in the WMT20 Metrics Shared Task
• Learning to Evaluate Translation Beyond English: BLEURT Submissions to the WMT Metrics 2020 Shared Task
• Designing Precise and Robust Dialogue Response Evaluators
• Naturalness Evaluation of Natural Language Generation in Task-oriented Dialogues using BERT
• Towards Automatic Evaluation of Dialog Systems: A Model-Free Off-Policy Evaluation Approach