Unanswerable Questions about Images and Texts

TL;DR

This paper discusses the challenges and issues related to unanswerable questions in visual and textual question answering systems, highlighting their significance for AI development.

Contribution

It provides an analysis of unanswerable questions across different AI tasks, emphasizing their importance and the need for better handling in AI models.

Findings

Unanswerable questions pose unique challenges in VQA and QA.

Addressing unanswerable questions is crucial for advancing AI robustness.

The paper highlights the need for improved methods to detect unanswerable queries.

Abstract

Questions about a text or an image that cannot be answered raise distinctive issues for an AI. This note discusses the problem of unanswerable questions in VQA (visual question answering), in QA (visual question answering), and in AI generally.