Abstract
In this paper, we address the hallucination problem commonly found in natural language generation tasks. Language models often generate fluent and convincing content but can lack consistency with the provided source, resulting in potential inaccuracies. We propose a new decoding method called Fidelity-Enriched Contrastive Search (FECS), which augments the contrastive search framework with context-aware regularization terms. FECS promotes tokens that are semantically similar to the provided source while penalizing repetitiveness in the generated text. We demonstrate its effectiveness across two tasks prone to hallucination: abstractive summarization and dialogue generation. Results show that FECS consistently enhances faithfulness across various language model sizes while maintaining output diversity comparable to well-performing decoding algorithms.- Anthology ID:
- 2023.emnlp-main.54
- 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:
- 843–851
- Language:
- URL:
- https://aclanthology.org/2023.emnlp-main.54
- DOI:
- 10.18653/v1/2023.emnlp-main.54
- Cite (ACL):
- Wei-Lin Chen, Cheng-Kuang Wu, Hsin-Hsi Chen, and Chung-Chi Chen. 2023. Fidelity-Enriched Contrastive Search: Reconciling the Faithfulness-Diversity Trade-Off in Text Generation. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 843–851, Singapore. Association for Computational Linguistics.
- Cite (Informal):
- Fidelity-Enriched Contrastive Search: Reconciling the Faithfulness-Diversity Trade-Off in Text Generation (Chen et al., EMNLP 2023)
- PDF:
- https://preview.aclanthology.org/nschneid-patch-4/2023.emnlp-main.54.pdf