In this lecture, the information-theoretical foundations of identity and privacy are provided. The framework is used to derive basic performance limits of access control systems, biometric authentication, privacy preserving data release mechanisms, generation of synthetic databases, etc. The lecture is a good supplement to lecture 19302801 "Applied Biometry".
The outline is as follows:
1) Introduction: Access control systems and biometrics, information-theoretic tools
2) Secret sharing with privacy constraints and impostors
3) Biometric authentication, zero leakage, generated and chosen secrets
4) Biometric identification, secret binding
5) Practical constraints, fuzzy commitment
6) Data bases, syntectic data, information-theoretic tools
7) Information-theoretic differential privacy
8) Laplace and Exponential Mechanism, Composition
9) Differential privacy and rate distotion, information bottleneck approach
10) Practical constraints, Google RAPPOR
11) Systems with learning (variational autoencoder concepts)