Exploring Prompt-based Few-shot Learning for Grounded Dialog Generation

TL;DR

This paper introduces a prompt-based approach for few-shot grounded dialog generation, demonstrating its effectiveness across multiple tasks and models, and providing insights into factors influencing prompt performance.

Contribution

It proposes a simple prompting method for grounded dialog generation and thoroughly analyzes its effectiveness across different models and tasks.

Findings

Prompted language models outperform conversational models in GDG tasks.

The prompting method's effectiveness varies with different pre-trained models.

Various factors significantly influence the success of prompting in GDG.

Abstract

Dialog models can be greatly strengthened through grounding on various external information, but grounded dialog corpora are usually not naturally accessible. In this work, we focus on the few-shot learning for grounded dialog generation (GDG). We first propose a simple prompting method for GDG tasks, where different constructs of model input, such as the grounding source and the conversation context, are distinguished through continuous or discrete prompts. On three typical GDG tasks, we empirically demonstrate and analyze in-depth the effectiveness of our method. We then conduct extensive experiments to thoroughly investigate how our prompting method works with different pre-trained models. We show that prompted language models perform superiorly to conversational models, and further analyze various factors that influence the effects of prompting. Overall, our work introduces a prompt-based perspective to the few-shot learning for GDG tasks, and provides valuable findings and insights for future research.