Gender and Racial Bias in Visual Question Answering Datasets

Abstract

Vision-and-language tasks have increasingly drawn more attention as a means to evaluate human-like reasoning in machine learning models. A popular task in the field is visual question answering (VQA), which aims to answer questions about images. However, VQA models have been shown to exploit language bias by learning the statistical correlations between questions and answers without looking into the image content: e.g., questions about the color of a banana are answered with yellow, even if the banana in the image is green. If societal bias (e.g., sexism, racism, ableism, etc.) is present in the training data, this problem may be causing VQA models to learn harmful stereotypes. For this reason, we investigate gender and racial bias in five VQA datasets. In our analysis, we find that the distribution of answers is highly different between questions about women and men, as well as the existence of detrimental gender-stereotypical samples. Likewise, we identify that specific race-related attributes are underrepresented, whereas potentially discriminatory samples appear in the analyzed datasets. Our findings suggest that there are dangers associated to using VQA datasets without considering and dealing with the potentially harmful stereotypes. We conclude the paper by proposing solutions to alleviate the problem before, during, and after the dataset collection process.

ACM Conference on Fairness, Accountability, and Transparency (FAccT 2022)

References in corpus (6)

• Microsoft COCO Captions: Data Collection and Evaluation Server
• Show and Tell: Lessons learned from the 2015 MSCOCO Image Captioning Challenge
• Towards a Critical Race Methodology in Algorithmic Fairness
• No Classification without Representation: Assessing Geodiversity Issues in Open Data Sets for the Developing World
• One Label, One Billion Faces: Usage and Consistency of Racial Categories in Computer Vision
• Feature and Label Embedding Spaces Matter in Addressing Image Classifier Bias

Cited by in corpus (3)

• From Image to Language: A Critical Analysis of Visual Question Answering (VQA) Approaches, Challenges, and Opportunities
• Graph Neural Networks in Vision-Language Image Understanding: A Survey
• Evaluating the Knowledge Dependency of Questions