In this paper, we address the limitations of existing text-to-image diffusion models in generating demographically fair results when given human-related descriptions. These models often struggle to disentangle the target language context from sociocultural biases, resulting in biased image generation. To overcome this challenge, we propose Fair Mapping, a flexible, model-agnostic, and lightweight approach that modifies a pre-trained text-to-image diffusion model by controlling the prompt to achieve fair image generation. One key advantage of our approach is its high efficiency. It only requires updating an additional linear network with few parameters at a low computational cost. By developing a linear network that maps conditioning embeddings into a debiased space, we enable the generation of relatively balanced demographic results based on the specified text condition. With comprehensive experiments on face image generation, we show that our method significantly improves image generation fairness with almost the same image quality compared to conventional diffusion models when prompted with descriptions related to humans. By effectively addressing the issue of implicit language bias, our method produces more fair and diverse image outputs.

Recently, federated learning (FL) has achieved wide successes for diverse privacy-sensitive applications without sacrificing the sensitive private information of clients. However, the data quality of client datasets can not be guaranteed since corresponding annotations of different clients often contain complex label noise of varying degrees, which inevitably causes the performance degradation. Intuitively, the performance degradation is dominated by clients with higher noise rates since their trained models contain more misinformation from data, thus it is necessary to devise an effective optimization scheme to mitigate the negative impacts of these noisy clients. In this work, we propose a two-stage framework FedELC to tackle this complicated label noise issue. The first stage aims to guide the detection of noisy clients with higher label noise, while the second stage aims to correct the labels of noisy clients' data via an end-to-end label correction framework which is achieved by learning possible ground-truth labels of noisy clients' datasets via back propagation. We implement sixteen related methods and evaluate five datasets with three types of complicated label noise scenarios for a comprehensive comparison. Extensive experimental results demonstrate our proposed framework achieves superior performance than its counterparts for different scenarios. Additionally, we effectively improve the data quality of detected noisy clients' local datasets with our label correction framework. The code is available at https://github.com/Sprinter1999/FedELC.

Rapid advancements in face recognition systems have revolutionized the field, enabling the efficiency and accuracy of various applications. Despite the significant progress achieved, these systems still face various challenges, with one important issue revolving around the potential biases and unfairness they may exhibit. These systems are likely to exhibit homogeneity bias, which refers to their tendency to erroneously classify faces belonging to the same gender or ethnic group together. Empirical observations indicate that the current representations of face data do not fully exploit the inherent dimension and expressive capacity of the representation space, leading to a diversity gap in representations for various groups. Consequently, under-utilized representation space for minority groups leads to bias in face recognition. In this paper, to enhance the representation space utilization and reduce disparities among different demographic groups, we introduce SA-SVD, a regularization method orthogonal to existing face recognition methods. It has the potential to capture the intricate patterns and subtle differences in face features across individuals within certain demographic groups. Extensive experiments demonstrate our method’s effectiveness in mitigating biased outcomes and achieving superior performance on two benchmark datasets VGGFace2 and CelebA.