BinaryVQA: A Versatile Test Set to Evaluate the Out-of-Distribution Generalization of VQA Models

TL;DR

BinaryVQA is a new binary question dataset designed to evaluate the out-of-distribution generalization of VQA models, revealing their limitations and biases through comprehensive analysis.

Contribution

The paper introduces BinaryVQA, a challenging, manually verified binary question dataset for VQA, and provides detailed analysis of model performance and biases.

Findings

OFA model achieves 75% accuracy on BinaryVQA, lower than on VQA v2.

BinaryVQA exposes model biases towards positive answers.

The dataset challenges VQA models and can guide future improvements.

Abstract

We introduce a new test set for visual question answering (VQA) called BinaryVQA to push the limits of VQA models. Our dataset includes 7,800 questions across 1,024 images and covers a wide variety of objects, topics, and concepts. For easy model evaluation, we only consider binary questions. Questions and answers are formulated and verified carefully and manually. Around 63% of the questions have positive answers. The median number of questions per image and question length are 7 and 5, respectively. The state of the art OFA model achieves 75% accuracy on BinaryVQA dataset, which is significantly lower than its performance on the VQA v2 test-dev dataset (94.7%). We also analyze the model behavior along several dimensions including: a) performance over different categories such as text, counting and gaze direction, b) model interpretability, c) the effect of question length on accuracy, d) bias of models towards positive answers and introduction of a new score called the ShuffleAcc, and e) sensitivity to spelling and grammar errors. Our investigation demonstrates the difficulty of our dataset and shows that it can challenge VQA models for next few years. Data and code are publicly available at: DATA and CODE.