Revealing and Reducing Gender Biases in Vision and Language Assistants (VLAs)

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Girrbach, Leander, Alaniz, Stephan, Huang, Yiran, Darrell, Trevor, Akata, Zeynep
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929756715352064
author Girrbach, Leander
Alaniz, Stephan
Huang, Yiran
Darrell, Trevor
Akata, Zeynep
author_facet Girrbach, Leander
Alaniz, Stephan
Huang, Yiran
Darrell, Trevor
Akata, Zeynep
contents Pre-trained large language models (LLMs) have been reliably integrated with visual input for multimodal tasks. The widespread adoption of instruction-tuned image-to-text vision-language assistants (VLAs) like LLaVA and InternVL necessitates evaluating gender biases. We study gender bias in 22 popular open-source VLAs with respect to personality traits, skills, and occupations. Our results show that VLAs replicate human biases likely present in the data, such as real-world occupational imbalances. Similarly, they tend to attribute more skills and positive personality traits to women than to men, and we see a consistent tendency to associate negative personality traits with men. To eliminate the gender bias in these models, we find that fine-tuning-based debiasing methods achieve the best trade-off between debiasing and retaining performance on downstream tasks. We argue for pre-deploying gender bias assessment in VLAs and motivate further development of debiasing strategies to ensure equitable societal outcomes. Code is available at https://github.com/ExplainableML/vla-gender-bias.
format Preprint
id arxiv_https___arxiv_org_abs_2410_19314
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Revealing and Reducing Gender Biases in Vision and Language Assistants (VLAs)
Girrbach, Leander
Alaniz, Stephan
Huang, Yiran
Darrell, Trevor
Akata, Zeynep
Computers and Society
Computation and Language
Pre-trained large language models (LLMs) have been reliably integrated with visual input for multimodal tasks. The widespread adoption of instruction-tuned image-to-text vision-language assistants (VLAs) like LLaVA and InternVL necessitates evaluating gender biases. We study gender bias in 22 popular open-source VLAs with respect to personality traits, skills, and occupations. Our results show that VLAs replicate human biases likely present in the data, such as real-world occupational imbalances. Similarly, they tend to attribute more skills and positive personality traits to women than to men, and we see a consistent tendency to associate negative personality traits with men. To eliminate the gender bias in these models, we find that fine-tuning-based debiasing methods achieve the best trade-off between debiasing and retaining performance on downstream tasks. We argue for pre-deploying gender bias assessment in VLAs and motivate further development of debiasing strategies to ensure equitable societal outcomes. Code is available at https://github.com/ExplainableML/vla-gender-bias.
title Revealing and Reducing Gender Biases in Vision and Language Assistants (VLAs)
topic Computers and Society
Computation and Language
url https://arxiv.org/abs/2410.19314