Abstract
Being able to rank the similarity of short text segments is an interesting bonus feature of neural machine translation. Translation-based similarity measures include direct and pivot translation probability, as well as translation cross-likelihood, which has not been studied so far. We analyze these measures in the common framework of multilingual NMT, releasing the NMTScore library. Compared to baselines such as sentence embeddings, translation-based measures prove competitive in paraphrase identification and are more robust against adversarial or multilingual input, especially if proper normalization is applied. When used for reference-based evaluation of data-to-text generation in 2 tasks and 17 languages, translation-based measures show a relatively high correlation to human judgments.- Anthology ID:
- 2022.findings-emnlp.15
- Volume:
- Findings of the Association for Computational Linguistics: EMNLP 2022
- Month:
- December
- Year:
- 2022
- Address:
- Abu Dhabi, United Arab Emirates
- Venue:
- Findings
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 198–213
- Language:
- URL:
- https://aclanthology.org/2022.findings-emnlp.15
- DOI:
- Cite (ACL):
- Jannis Vamvas and Rico Sennrich. 2022. NMTScore: A Multilingual Analysis of Translation-based Text Similarity Measures. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 198–213, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
- Cite (Informal):
- NMTScore: A Multilingual Analysis of Translation-based Text Similarity Measures (Vamvas & Sennrich, Findings 2022)
- PDF:
- https://preview.aclanthology.org/ingestion-script-update/2022.findings-emnlp.15.pdf