KETOD: Knowledge-Enriched Task-Oriented Dialogue

TL;DR

This paper introduces KETOD, a new dataset and models that integrate task-oriented dialogue with knowledge-grounded chit-chat, enhancing response quality while maintaining task performance.

Contribution

The work presents a novel dataset, KETOD, and two models, SimpleToDPlus and Combiner, for unified task-oriented and knowledge-enriched chit-chat dialogue systems.

Findings

Models significantly improve knowledge-enriched response generation.

Maintains competitive task-oriented dialogue performance.

Dataset and code are publicly available.

Abstract

Existing studies in dialogue system research mostly treat task-oriented dialogue and chit-chat as separate domains. Towards building a human-like assistant that can converse naturally and seamlessly with users, it is important to build a dialogue system that conducts both types of conversations effectively. In this work, we investigate how task-oriented dialogue and knowledge-grounded chit-chat can be effectively integrated into a single model. To this end, we create a new dataset, KETOD (Knowledge-Enriched Task-Oriented Dialogue), where we naturally enrich task-oriented dialogues with chit-chat based on relevant entity knowledge. We also propose two new models, SimpleToDPlus and Combiner, for the proposed task. Experimental results on both automatic and human evaluations show that the proposed methods can significantly improve the performance in knowledge-enriched response generation while maintaining a competitive task-oriented dialog performance. We believe our new dataset will be a valuable resource for future studies. Our dataset and code are publicly available at \url{https://github.com/facebookresearch/ketod}.