arXiv subjects

Callum Mole

Publications and source records attributed to Callum Mole.

4 recordsLinked to original sources

arXiv2023

Digital identity architectures: comparing goals and vulnerabilities

Digital identity systems have the promise of efficiently facilitating access to services for a nation's citizens while increasing security and convenience. There are many possible system architectures, each with strengths and weaknesses that should be carefully considered. This report first establishes a set of goals and vulnerabilities faced by any identity system, then evaluates the trade-offs of common digital identity architectures, principally comparing centralised and decentralised systems.

cs.CR↗

arXiv2022

A Framework for Auditable Synthetic Data Generation

Synthetic data has gained significant momentum thanks to sophisticated machine learning tools that enable the synthesis of high-dimensional datasets. However, many generation techniques do not give the data controller control over what statistical patterns are captured, leading to concerns over privacy protection. While synthetic records are not linked to a particular real-world individual, they can reveal information about users indirectly which may be unacceptable for data owners. There is thus a need to empirically verify the privacy of synthetic data -- a particularly challenging task in high-dimensional data. In this paper we present a general framework for synthetic data generation that gives data controllers full control over which statistical properties the synthetic data ought to preserve, what exact information loss is acceptable, and how to quantify it. The benefits of the approach are that (1) one can generate synthetic data that results in high utility for a given task, while (2) empirically validating that only statistics considered safe by the data curator are used to generate the data. We thus show the potential for synthetic data to be an effective means of releasing confidential data safely, while retaining useful information for analysts.

cs.CR↗

arXiv2022

TAPAS: a Toolbox for Adversarial Privacy Auditing of Synthetic Data

Personal data collected at scale promises to improve decision-making and accelerate innovation. However, sharing and using such data raises serious privacy concerns. A promising solution is to produce synthetic data, artificial records to share instead of real data. Since synthetic records are not linked to real persons, this intuitively prevents classical re-identification attacks. However, this is insufficient to protect privacy. We here present TAPAS, a toolbox of attacks to evaluate synthetic data privacy under a wide range of scenarios. These attacks include generalizations of prior works and novel attacks. We also introduce a general framework for reasoning about privacy threats to synthetic data and showcase TAPAS on several examples.

cs.CR↗

arXiv2018

Modelling visual-vestibular integration and behavioural adaptation in the driving simulator

It is well established that not only vision but also other sensory modalities affect drivers' control of their vehicles, and that drivers adapt over time to persistent changes in sensory cues (for example in driving simulators), but the mechanisms underlying these behavioural phenomena are poorly understood. Here, we consider the existing literature on how driver steering in slalom tasks is affected by the down-scaling of vestibular cues, and propose a driver model that can explain the empirically observed effects, namely: decreased task performance and increased steering effort during initial exposure, followed by a partial reversal of these effects as task exposure is prolonged. Unexpectedly, the model also reproduced another empirical finding: a local optimum for motion down-scaling, where path-tracking is better than when one-to-one motion cues are available. Overall, the results imply that: (1) drivers make direct use of vestibular information as part of determining appropriate steering, and (2) motion down-scaling causes a yaw rate underestimation phenomenon, where drivers behave as if the simulated vehicle is rotating more slowly than it is. However, (3) in the slalom task, a certain degree of such yaw rate underestimation is beneficial to path tracking performance. Furthermore, (4) behavioural adaptation, as empirically observed in slalom tasks, may occur due to (a) down-weighting of vestibular cues, and/or (b) increased sensitivity to control errors, in determining when to adjust steering and by how much, but (c) seemingly not in the form of a full compensatory rescaling of the received vestibular input. The analyses presented here provide new insights and hypotheses about simulator driving, and the developed models can be used to support research on multisensory integration and behavioural adaptation in both driving and other task domains.

q-bio.NC↗