Adversary-Aware Private Inference over Wireless Channels
This addresses privacy concerns in edge AI applications like autonomous driving and environmental monitoring, but it appears incremental as it builds on existing differential privacy mechanisms for datasets.
The paper tackles the problem of protecting sensitive personal data during AI-based sensing at wireless edge devices by proposing a novel framework for privacy-preserving inference, where devices transform extracted features before transmission to a model server to reduce privacy risks from adversaries.
AI-based sensing at wireless edge devices has the potential to significantly enhance Artificial Intelligence (AI) applications, particularly for vision and perception tasks such as in autonomous driving and environmental monitoring. AI systems rely both on efficient model learning and inference. In the inference phase, features extracted from sensing data are utilized for prediction tasks (e.g., classification or regression). In edge networks, sensors and model servers are often not co-located, which requires communication of features. As sensitive personal data can be reconstructed by an adversary, transformation of the features are required to reduce the risk of privacy violations. While differential privacy mechanisms provide a means of protecting finite datasets, protection of individual features has not been addressed. In this paper, we propose a novel framework for privacy-preserving AI-based sensing, where devices apply transformations of extracted features before transmission to a model server.