Bias Detection and Rotation-Robustness Mitigation in Vision-Language Models and Generative Image Models

TL;DR

This paper investigates how image rotation affects the robustness and bias in vision-language and generative models, proposing mitigation strategies that improve fairness and stability under such transformations.

Contribution

It introduces rotation-robust mitigation techniques combining data augmentation, representation alignment, and regularization to enhance model robustness and fairness.

Findings

Significant improvement in robustness against image rotation.

Reduction in bias amplification without performance loss.

Effective mitigation strategies applicable across multiple datasets.

Abstract

Vision-Language Models (VLMs) and generative image models have achieved remarkable performance across multimodal tasks, yet their robustness and fairness under input transformations remain insufficiently explored. This work investigates bias propagation and robustness degradation in state-of-the-art vision-language and generative models, with a particular focus on image rotation and distributional shifts. We analyze how rotation-induced perturbations affect model predictions, confidence calibration, and demographic bias patterns. To address these issues, we propose rotation-robust mitigation strategies that combine data augmentation, representation alignment, and model-level regularization. Experimental results across multiple datasets demonstrate that the proposed methods significantly improve robustness while reducing bias amplification without sacrificing overall performance. This study highlights critical limitations of current multimodal systems and provides practical mitigation techniques for building more reliable and fair AI models.