Training privacy-preserving video analytics pipelines by suppressing features that reveal information about private attributes
This addresses privacy concerns in applications like out-of-home advertising analytics, but it is incremental as it builds on existing methods for privacy-preserving machine learning.
The paper tackles the problem of private attribute leakage in deep neural networks used for video analytics by modifying training with a confusion loss, reducing gender leakage by 2.88% and age group leakage by 13.06% while minimally affecting task accuracy.
Deep neural networks are increasingly deployed for scene analytics, including to evaluate the attention and reaction of people exposed to out-of-home advertisements. However, the features extracted by a deep neural network that was trained to predict a specific, consensual attribute (e.g. emotion) may also encode and thus reveal information about private, protected attributes (e.g. age or gender). In this work, we focus on such leakage of private information at inference time. We consider an adversary with access to the features extracted by the layers of a deployed neural network and use these features to predict private attributes. To prevent the success of such an attack, we modify the training of the network using a confusion loss that encourages the extraction of features that make it difficult for the adversary to accurately predict private attributes. We validate this training approach on image-based tasks using a publicly available dataset. Results show that, compared to the original network, the proposed PrivateNet can reduce the leakage of private information of a state-of-the-art emotion recognition classifier by 2.88% for gender and by 13.06% for age group, with a minimal effect on task accuracy.