Transformation Driven Visual Reasoning

TL;DR

This paper introduces a transformation driven visual reasoning paradigm and dataset, emphasizing the importance of inferring dynamic changes between states, which challenges current static reasoning models.

Contribution

It proposes a new reasoning task and dataset focusing on transformations between states, extending beyond static concept understanding in visual reasoning.

Findings

Models excel on single-step transformations but struggle with multi-step and view-invariant transformations.

The dataset TRANCE enables evaluation of dynamic reasoning capabilities.

Current models are far from human-level performance on complex transformation tasks.

Abstract

This paper defines a new visual reasoning paradigm by introducing an important factor, i.e.~transformation. The motivation comes from the fact that most existing visual reasoning tasks, such as CLEVR in VQA, are solely defined to test how well the machine understands the concepts and relations within static settings, like one image. We argue that this kind of \textbf{state driven visual reasoning} approach has limitations in reflecting whether the machine has the ability to infer the dynamics between different states, which has been shown as important as state-level reasoning for human cognition in Piaget's theory. To tackle this problem, we propose a novel \textbf{transformation driven visual reasoning} task. Given both the initial and final states, the target is to infer the corresponding single-step or multi-step transformation, represented as a triplet (object, attribute, value) or a sequence of triplets, respectively. Following this definition, a new dataset namely TRANCE is constructed on the basis of CLEVR, including three levels of settings, i.e.~Basic (single-step transformation), Event (multi-step transformation), and View (multi-step transformation with variant views). Experimental results show that the state-of-the-art visual reasoning models perform well on Basic, but are still far from human-level intelligence on Event and View. We believe the proposed new paradigm will boost the development of machine visual reasoning. More advanced methods and real data need to be investigated in this direction. The resource of TVR is available at https://hongxin2019.github.io/TVR.