Lecturer	Graham Cormode
Degrees	Schedule C1 (CS&P) — Computer Science and Philosophy Schedule C1 — Computer Science Schedule C1 — Mathematics and Computer Science Hilary Term — MSc in Advanced Computer Science
Term	Hilary Term 2027  (20 lectures)

Overview

Modern computer science applications make extensive use of data to offer recommendations, perform analytics and build models. Such data is often drawn from the characteristics, preferences and actions of individuals, meaning that extra care is needed in handling this data to comply with privacy legislation and provide consumer confidence. “Privacy Enhancing Technologies”, or PETs, form a key part of the privacy picture, offering technical controls to allow us to work with private data while preserving the privacy of the individuals who contribute to the data. This course takes a computational perspective on privacy. After setting the scene with the legal and social demands for privacy, the bulk of the coverage is on algorithms and tools that promise some notion of privacy. A major focus is on the model of Differential Privacy, which has been widely adopted for data analysis by large companies and national statistics agencies. Differential privacy introduces carefully calibrated statistical noise into computation results to create uncertainty around the contribution of any individual. The second half of the course will study other techniques that offer different notions of privacy, and assume different levels of trust. These may include Zero-knowledge proofs, Multiparty Computation, Synthetic Data Generation, Federated Learning, and Homomorphic Encryption.

Learning outcomes

By the end of this course, students should be able to:

• Explain the legal, social, and technical motivations for privacy in modern data-driven systems, and clearly distinguish privacy goals from traditional security goals.

• Analyse common privacy attacks and failures of naïve data anonymization techniques, including reidentification and inference attacks.

• Formally define differential privacy and apply its core mechanisms, including noise addition, composition, and explain utility-privacy trade-offs.

• Design and analyse differentially private algorithms for basic data analysis and machine learning tasks, including histograms, optimization, and model training.

• Understand and contrast variations of differential privacy and alternative trust models, such as local, distributed, and shuffled differential privacy.

• Explain the principles behind major privacy-enhancing technologies, including secure multi-party computation, federated learning, synthetic data generation, and zero-knowledge proofs.

• Assess the suitability, assumptions, and performance trade-offs of different privacy-preserving computation techniques for practical applications.

Taking our courses

This form is not to be used by students studying for a degree in the Department of Computer Science, or for Visiting Students who are registered for Computer Science courses

Other matriculated University of Oxford students who are interested in taking this, or other, courses in the Department of Computer Science, must complete this online form by 17.00 on Friday of 0th week of term in which the course is taught. Late requests, and requests sent by email, will not be considered. All requests must be approved by the relevant Computer Science departmental committee and can only be submitted using this form.