Fidelity-Enriched Contrastive Search: Reconciling the Faithfulness-Diversity Trade-Off in Text Generation
Journal
EMNLP 2023 - 2023 Conference on Empirical Methods in Natural Language Processing, Proceedings
ISBN
9798891760608
Date Issued
2023-01-01
Author(s)
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 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.
Type
conference paper