Enhancing Task Bot Engagement with Synthesized Open-Domain Dialog

TL;DR

This paper introduces PivotBot, a unified dialogue system capable of seamlessly switching between task-oriented and open-domain conversations by generating fused dialogues and accessing diverse knowledge sources.

Contribution

It proposes a novel framework for automatic generation of fused TOD and ODD dialogues and introduces PivotBot, a model that effectively handles both modes and knowledge sources.

Findings

PivotBot outperforms existing models in switching between TOD and ODD.

The framework successfully generates diverse, knowledge-grounded fused dialogues.

PivotBot demonstrates superior adaptability in mixed conversation settings.

Abstract

Many efforts have been made to construct dialog systems for different types of conversations, such as task-oriented dialog (TOD) and open-domain dialog (ODD). To better mimic human-level conversations that usually fuse various dialog modes, it is essential to build a system that can effectively handle both TOD and ODD and access different knowledge sources. To address the lack of available data for the fused task, we propose a framework for automatically generating dialogues that combine knowledge-grounded ODDs and TODs in various settings. Additionally, we introduce a unified model PivotBot that is capable of appropriately adopting TOD and ODD modes and accessing different knowledge sources in order to effectively tackle the fused task. Evaluation results demonstrate the superior ability of the proposed model to switch seamlessly between TOD and ODD tasks.