Ask No More: Deciding when to guess in referential visual dialogue

TL;DR

This paper introduces a decision-making component to visual dialogue agents, enabling them to decide when to ask questions or make guesses, resulting in more efficient and natural interactions.

Contribution

It presents a novel approach to incorporate decision-making into end-to-end visual dialogue models, improving dialogue quality and efficiency.

Findings

Fewer unnecessary questions in dialogues

Less repetitive dialogue interactions

Enhanced efficiency and naturalness

Abstract

Our goal is to explore how the abilities brought in by a dialogue manager can be included in end-to-end visually grounded conversational agents. We make initial steps towards this general goal by augmenting a task-oriented visual dialogue model with a decision-making component that decides whether to ask a follow-up question to identify a target referent in an image, or to stop the conversation to make a guess. Our analyses show that adding a decision making component produces dialogues that are less repetitive and that include fewer unnecessary questions, thus potentially leading to more efficient and less unnatural interactions.