Assessing Intersectional Bias in Representations of Pre-Trained Image Recognition Models
This work addresses fairness issues in AI for facial recognition systems, though it is incremental as it applies existing bias assessment methods to new intersectional contexts.
The researchers investigated intersectional biases in pre-trained ImageNet classifiers for facial images across age, race, and gender, finding that representations strongly differentiate ages and moderately associate ethnicities and distinguish genders in middle-aged groups.
Deep Learning models have achieved remarkable success. Training them is often accelerated by building on top of pre-trained models which poses the risk of perpetuating encoded biases. Here, we investigate biases in the representations of commonly used ImageNet classifiers for facial images while considering intersections of sensitive variables age, race and gender. To assess the biases, we use linear classifier probes and visualize activations as topographic maps. We find that representations in ImageNet classifiers particularly allow differentiation between ages. Less strongly pronounced, the models appear to associate certain ethnicities and distinguish genders in middle-aged groups.