Abstract
We introduce ParaBLEU, a paraphrase representation learning model and evaluation metric for text generation. Unlike previous approaches, ParaBLEU learns to understand paraphrasis using generative conditioning as a pretraining objective. ParaBLEU correlates more strongly with human judgements than existing metrics, obtaining new state-of-the-art results on the 2017 WMT Metrics Shared Task. We show that our model is robust to data scarcity, exceeding previous state-of-the-art performance using only 50% of the available training data and surpassing BLEU, ROUGE and METEOR with only 40 labelled examples. Finally, we demonstrate that ParaBLEU can be used to conditionally generate novel paraphrases from a single demonstration, which we use to confirm our hypothesis that it learns abstract, generalized paraphrase representations.- Anthology ID:
- 2022.acl-long.280
- Volume:
- Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Month:
- May
- Year:
- 2022
- Address:
- Dublin, Ireland
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 4052–4073
- Language:
- URL:
- https://aclanthology.org/2022.acl-long.280
- DOI:
- 10.18653/v1/2022.acl-long.280
- Cite (ACL):
- Jack Weston, Raphael Lenain, Udeepa Meepegama, and Emil Fristed. 2022. Generative Pretraining for Paraphrase Evaluation. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 4052–4073, Dublin, Ireland. Association for Computational Linguistics.
- Cite (Informal):
- Generative Pretraining for Paraphrase Evaluation (Weston et al., ACL 2022)
- PDF:
- https://preview.aclanthology.org/ingestion-script-update/2022.acl-long.280.pdf
- Data
- GLUE, MRPC, MultiNLI, PARANMT-50M, PAWS, SNLI