Gone With the Bits: Revealing Racial Bias in Low-Rate Neural Compression for Facial Images
This addresses fairness concerns in neural compression for facial images, which is important for applications like surveillance and social media, though it's an incremental step in bias evaluation.
The paper investigates racial bias in neural image compression models, finding that all nine popular models tested exhibit bias that can be captured through facial phenotype degradation in reconstructions, and shows that using racially balanced training data reduces but doesn't eliminate bias.
Neural compression methods are gaining popularity due to their superior rate-distortion performance over traditional methods, even at extremely low bitrates below 0.1 bpp. As deep learning architectures, these models are prone to bias during the training process, potentially leading to unfair outcomes for individuals in different groups. In this paper, we present a general, structured, scalable framework for evaluating bias in neural image compression models. Using this framework, we investigate racial bias in neural compression algorithms by analyzing nine popular models and their variants. Through this investigation, we first demonstrate that traditional distortion metrics are ineffective in capturing bias in neural compression models. Next, we highlight that racial bias is present in all neural compression models and can be captured by examining facial phenotype degradation in image reconstructions. We then examine the relationship between bias and realism in the decoded images and demonstrate a trade-off across models. Finally, we show that utilizing a racially balanced training set can reduce bias but is not a sufficient bias mitigation strategy. We additionally show the bias can be attributed to compression model bias and classification model bias. We believe that this work is a first step towards evaluating and eliminating bias in neural image compression models.