Fighting Selection Bias in Statistical Learning: Application to Visual Recognition from Biased Image Databases
This addresses representativeness issues in facial recognition systems, which can have uneven performances across population segments, and is incremental as it builds on existing biasing models.
The paper tackles the problem of selection bias in visual recognition by proposing a method to reweight observations from biased image databases, resulting in a nearly debiased estimator of the target distribution.
In practice, and especially when training deep neural networks, visual recognition rules are often learned based on various sources of information. On the other hand, the recent deployment of facial recognition systems with uneven performances on different population segments has highlighted the representativeness issues induced by a naive aggregation of the datasets. In this paper, we show how biasing models can remedy these problems. Based on the (approximate) knowledge of the biasing mechanisms at work, our approach consists in reweighting the observations, so as to form a nearly debiased estimator of the target distribution. One key condition is that the supports of the biased distributions must partly overlap, and cover the support of the target distribution. In order to meet this requirement in practice, we propose to use a low dimensional image representation, shared across the image databases. Finally, we provide numerical experiments highlighting the relevance of our approach.