Bounding the Invertibility of Privacy-preserving Instance Encoding using Fisher Information
This addresses the need for rigorous privacy assessment in machine learning for data protection, though it is incremental as it builds on existing encoding methods.
The paper tackled the problem of evaluating the privacy of instance encoding schemes by proposing a theoretically-principled measure based on Fisher information, showing it can bound the invertibility of encodings both theoretically and empirically.
Privacy-preserving instance encoding aims to encode raw data as feature vectors without revealing their privacy-sensitive information. When designed properly, these encodings can be used for downstream ML applications such as training and inference with limited privacy risk. However, the vast majority of existing instance encoding schemes are based on heuristics and their privacy-preserving properties are only validated empirically against a limited set of attacks. In this paper, we propose a theoretically-principled measure for the privacy of instance encoding based on Fisher information. We show that our privacy measure is intuitive, easily applicable, and can be used to bound the invertibility of encodings both theoretically and empirically.