Abstract
Automatic evaluation for sentence simplification remains a challenging problem. Most popular evaluation metrics require multiple high-quality references – something not readily available for simplification – which makes it difficult to test performance on unseen domains. Furthermore, most existing metrics conflate simplicity with correlated attributes such as fluency or meaning preservation. We propose a new learned evaluation metric — SLE — which focuses on simplicity, outperforming almost all existing metrics in terms of correlation with human judgements.- Anthology ID:
- 2023.emnlp-main.739
- Volume:
- Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing
- Month:
- December
- Year:
- 2023
- Address:
- Singapore
- Editors:
- Houda Bouamor, Juan Pino, Kalika Bali
- Venue:
- EMNLP
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 12053–12059
- Language:
- URL:
- https://aclanthology.org/2023.emnlp-main.739
- DOI:
- 10.18653/v1/2023.emnlp-main.739
- Cite (ACL):
- Liam Cripwell, Joël Legrand, and Claire Gardent. 2023. Simplicity Level Estimate (SLE): A Learned Reference-Less Metric for Sentence Simplification. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 12053–12059, Singapore. Association for Computational Linguistics.
- Cite (Informal):
- Simplicity Level Estimate (SLE): A Learned Reference-Less Metric for Sentence Simplification (Cripwell et al., EMNLP 2023)
- PDF:
- https://preview.aclanthology.org/corrections-2024-05/2023.emnlp-main.739.pdf