CLEVR_HYP: A Challenge Dataset and Baselines for Visual Question Answering with Hypothetical Actions over Images

TL;DR

This paper introduces CLEVR_HYP, a new dataset and baseline models for visual question answering involving hypothetical actions, challenging systems to reason beyond explicit image content by simulating potential outcomes.

Contribution

It formulates a novel VQA task involving hypothetical reasoning, creates a corresponding dataset, and adapts existing models to benchmark this complex visual understanding challenge.

Findings

Baseline models demonstrate limited reasoning capabilities on hypothetical scenarios.

The dataset enables evaluation of joint reasoning over images and language.

Insights into architecture performance for complex visual reasoning tasks.

Abstract

Most existing research on visual question answering (VQA) is limited to information explicitly present in an image or a video. In this paper, we take visual understanding to a higher level where systems are challenged to answer questions that involve mentally simulating the hypothetical consequences of performing specific actions in a given scenario. Towards that end, we formulate a vision-language question answering task based on the CLEVR (Johnson et. al., 2017) dataset. We then modify the best existing VQA methods and propose baseline solvers for this task. Finally, we motivate the development of better vision-language models by providing insights about the capability of diverse architectures to perform joint reasoning over image-text modality. Our dataset setup scripts and codes will be made publicly available at https://github.com/shailaja183/clevr_hyp.