Source author record

Giacomo Dossena

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record

math-ph math.MP Cryptography and Security Machine Learning math.AT math.GN

Catalog footprint

Research overview

5works

6topics

4close collaborators

Actions

Is this your work?

Open graph Browse papers

preprint2022arXiv

Assessing the risk of re-identification arising from an attack on anonymised data

Objective: The use of routinely-acquired medical data for research purposes requires the protection of patient confidentiality via data anonymisation. The objective of this work is to calculate the risk of re-identification arising from a malicious attack to an anonymised dataset, as described below. Methods: We first present an analytical means of estimating the probability of re-identification of a single patient in a k-anonymised dataset of Electronic Health Record (EHR) data. Second, we generalize this solution to obtain the probability of multiple patients being re-identified. We provide synthetic validation via Monte Carlo simulations to illustrate the accuracy of the estimates obtained. Results: The proposed analytical framework for risk estimation provides re-identification probabilities that are in agreement with those provided by simulation in a number of scenarios. Our work is limited by conservative assumptions which inflate the re-identification probability. Discussion: Our estimates show that the re-identification probability increases with the proportion of the dataset maliciously obtained and that it has an inverse relationship with the equivalence class size. Our recursive approach extends the applicability domain to the general case of a multi-patient re-identification attack in an arbitrary k-anonymisation scheme. Conclusion: We prescribe a systematic way to parametrize the k-anonymisation process based on a pre-determined re-identification probability. We observed that the benefits of a reduced re-identification risk that come with increasing k-size may not be worth the reduction in data granularity when one is considering benchmarking the re-identification probability on the size of the portion of the dataset maliciously obtained by the adversary.

preprint2013arXiv

Sylvester's law of inertia for quadratic forms on vector bundles

This paper presents a generalisation of Sylvester's law of inertia to real non-degenerate quadratic forms on a fixed real vector bundle over a connected locally connected paracompact Hausdorff space. By interpreting the classical inertia as a complete discrete invariant for the natural action of the general linear group on quadratic forms, the simplest generalisation consists in substituting such group with the group of gauge transformations of the bundle. Contrary to the classical law of inertia, here the full action and its restriction to the identity path component typically have different orbits, leading to two invariants: a complete invariant for the full action is given by the isomorphism class of the orthonormal frame bundle associated to a quadratic form, while a complete invariant for the restricted action is the homotopy class of any maximal positive-definite subbundle associated to a quadratic form. The latter invariant is finer than the former, which in turn is finer than inertia. Moreover, the orbit structure thus obtained might be used to shed light on the topology of the space of non-degenerate quadratic forms on a vector bundle.

preprint2012arXiv

Dirac operator on spinors and diffeomorphisms

The issue of general covariance of spinors and related objects is reconsidered. Given an oriented manifold $M$ , to each spin structure $σ$ and Riemannian metric $g$ there is associated a space $S_{σ, g}$ of spinor fields on $M$ and a Hilbert space $\HH_{σ, g}= L^2(S_{σ, g},\vol{M}{g})$ of $L^2$ -spinors of $S_{σ, g}$ . The group $\diff{M}$ of orientation-preserving diffeomorphisms of $M$ acts both on $g$ (by pullback) and on $[σ]$ (by a suitably defined pullback $f^*σ$ ). Any $f\in \diff{M}$ lifts in exactly two ways to a unitary operator $U$ from $\HH_{σ, g} $ to $\HH_{f^*σ,f^*g}$. The canonically defined Dirac operator is shown to be equivariant with respect to the action of $U$ , so in particular its spectrum is invariant under the diffeomorphisms.

preprint2011arXiv

Comment on the paper "Does Zeeman's Fine Topology Exist?" at arXiv:1003.3703v1

A constructive and straightforward proof of the existence of the Zeeman topology is provided, contradicting a fallacious claim contained in the paper "Does Zeeman's Fine Topology Exist?" available at arXiv:1003.3703v1.

preprint2010arXiv

Product of real spectral triples

We construct the product of real spectral triples of arbitrary finite dimension (and arbitrary parity) taking into account the fact that in the even case there are two possible real structures, in the odd case there are two inequivalent representations of the gamma matrices (Clifford algebra), and in the even-even case there are two natural candidates for the Dirac operator of the product triple.

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.