RHO ($\rho$): Reducing Hallucination in Open-domain Dialogues with Knowledge Grounding

TL;DR

This paper introduces RHO, a knowledge-grounded dialogue system that reduces hallucinations by integrating linked entity and relation representations from a knowledge graph, employing local and global grounding, and a re-ranking technique, achieving significant improvements.

Contribution

RHO combines local and global knowledge grounding with a novel re-ranking method to enhance faithfulness in dialogue responses, addressing hallucination issues in knowledge-grounded systems.

Findings

Significantly reduces hallucination in responses by 17.54% in FeQA.

Outperforms state-of-the-art methods on OpenDialKG.

Improves both automatic and human evaluation metrics.

Abstract

Dialogue systems can leverage large pre-trained language models and knowledge to generate fluent and informative responses. However, these models are still prone to produce hallucinated responses not supported by the input source, which greatly hinders their application. The heterogeneity between external knowledge and dialogue context challenges representation learning and source integration, and further contributes to unfaithfulness. To handle this challenge and generate more faithful responses, this paper presents RHO ($\rho$) utilizing the representations of linked entities and relation predicates from a knowledge graph (KG). We propose (1) local knowledge grounding to combine textual embeddings with the corresponding KG embeddings; and (2) global knowledge grounding to equip RHO with multi-hop reasoning abilities via the attention mechanism. In addition, we devise a response re-ranking technique based on walks over KG sub-graphs for better conversational reasoning. Experimental results on OpenDialKG show that our approach significantly outperforms state-of-the-art methods on both automatic and human evaluation by a large margin, especially in hallucination reduction (17.54% in FeQA).