Adam James Hall

Nick Angelou, Ayoub Benaissa, Bogdan Cebere et al.

We present a multi-language, cross-platform, open-source library for asymmetric private set intersection (PSI) and PSI-Cardinality (PSI-C). Our protocol combines traditional DDH-based PSI and PSI-C protocols with compression based on Bloom filters that helps reduce communication in the asymmetric setting. Currently, our library supports C++, C, Go, WebAssembly, JavaScript, Python, and Rust, and runs on both traditional hardware (x86) and browser targets. We further apply our library to two use cases: (i) a privacy-preserving contact tracing protocol that is compatible with existing approaches, but improves their privacy guarantees, and (ii) privacy-preserving machine learning on vertically partitioned data.

Adam James Hall, Madhava Jay, Tudor Cebere et al.

We present Syft 0.5, a general-purpose framework that combines a core group of privacy-enhancing technologies that facilitate a universal set of structured transparency systems. This framework is demonstrated through the design and implementation of a novel privacy-preserving inference information flow where we pass homomorphically encrypted activation signals through a split neural network for inference. We show that splitting the model further up the computation chain significantly reduces the computation time of inference and the payload size of activation signals at the cost of model secrecy. We evaluate our proposed flow with respect to its provision of the core structural transparency principles.

Tom Titcombe, Adam J. Hall, Pavlos Papadopoulos et al.

We describe a threat model under which a split network-based federated learning system is susceptible to a model inversion attack by a malicious computational server. We demonstrate that the attack can be successfully performed with limited knowledge of the data distribution by the attacker. We propose a simple additive noise method to defend against model inversion, finding that the method can significantly reduce attack efficacy at an acceptable accuracy trade-off on MNIST. Furthermore, we show that NoPeekNN, an existing defensive method, protects different information from exposure, suggesting that a combined defence is necessary to fully protect private user data.

Daniele Romanini, Adam James Hall, Pavlos Papadopoulos et al.

We introduce PyVertical, a framework supporting vertical federated learning using split neural networks. The proposed framework allows a data scientist to train neural networks on data features vertically partitioned across multiple owners while keeping raw data on an owner's device. To link entities shared across different datasets' partitions, we use Private Set Intersection on IDs associated with data points. To demonstrate the validity of the proposed framework, we present the training of a simple dual-headed split neural network for a MNIST classification task, with data samples vertically distributed across two data owners and a data scientist.

Adam James Hall, Nikolaos Pitropakis, William J Buchanan et al.

Insider threats continue to present a major challenge for the information security community. Despite constant research taking place in this area; a substantial gap still exists between the requirements of this community and the solutions that are currently available. This paper uses the CERT dataset r4.2 along with a series of machine learning classifiers to predict the occurrence of a particular malicious insider threat scenario - the uploading sensitive information to wiki leaks before leaving the organization. These algorithms are aggregated into a meta-classifier which has a stronger predictive performance than its constituent models. It also defines a methodology for performing pre-processing on organizational log data into daily user summaries for classification, and is used to train multiple classifiers. Boosting is also applied to optimise classifier accuracy. Overall the models are evaluated through analysis of their associated confusion matrix and Receiver Operating Characteristic (ROC) curve, and the best performing classifiers are aggregated into an ensemble classifier. This meta-classifier has an accuracy of \textbf{96.2\%} with an area under the ROC curve of \textbf{0.988}.