Confidentialité différentielle

Résumé

La confidentialité différentielle, en bases de données et parfois associé à la cryptographie, est une propriété d'anonymisation pouvant être atteinte via différents mécanismes. Celle-ci vise à définir une forme de protection de résultats de requêtes faites à une de bases de données en minimisant les risques d'identification des entités qu'elle contient, si possible en maximisant la pertinence des résultats de la requête. Elle permet notamment l'exploitation statistique de données individuelles agrégées, sans compromettre la vie privée des individus. Motivations Généralités La notion d'intimité différentielle est une formalisation du concept intuitif de confidentialité dans le contexte des bases de données statistiques. Un tiers de confiance dépositaire d'une base de données d'informations sensibles (par exemple dossiers médicaux, registres d'électeurs, métadonnées de courriels) peut souhaiter fournir des informations statistiques, globales sur ces données. C

Source officielle

https://fr.wikipedia.org/wiki/Confidentialit%C3%A9_diff%C3%A9rentielle

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Publications associées (33)

Publications associées

Chargement

Personnes associées (6)

Personnes associées

Chargement

Unités associées (7)

Unités associées

Chargement

Concepts associés (1)

Concepts associés

Chargement

Cours associés (11)

Cours associés

Chargement

Séances de cours associées (40)

Séances de cours associées

Chargement

The strength of the Barcelona's Eixample comes from its urban plan, a visionary project designed by Ildefons Cerdà in 1859. The original project proposed blocks with a central open space built on two or three sides, allowing the introduction of variation through the urban fabric. However, to increase density the courtyards were privatized thereby reducing the public realm to the streets. The process was nevertheless reversed in 1986 by a new bylaw promoting the recuperation of the courtyards for public space. This project reasserts Cerdà's agenda for a diverse public realm forming a fluid network through streets and reclaimed courtyards and creating unique localities along promenades within a repetitive system. To solve the relation between users and programs, the different uses can coexist by a set of levels, taking up the existing fragmentation of buildings' terraces. Housing is extended by communal gardens, preserving privacy, while cafés and retail expand their activities on the interior and children and the elderly find social activities. The design of these itineraries is solved through parametric optimization, using a gaming pathfinder algorithm, with emphasis on the privacy condition of each block. The engine uses agents (virtual people) to determine the appropriate location and permeability of the threshold between street and courtyard. It is able to manage systematically small, human-scale interventions to (re) introduce a walking dimension into Cerdà's urban fabric.

2012

Chargement

Protecting Privacy in Multi-agent Optimization

Eric Zbinden

Many real-life optimization problems involve multiple entities, or agents (individuals, companies...), with their own private constraints and preferences, communicating with each other in order to find a solution that maximizes the overall public satisfaction. In Artificial Intelligence, the field of Distributed Constraint Optimization (DCOP) has been addressing such multi-agent optimization problems, through distributed message-passing algorithms such as the DPOP algorithm. However, the research in DCOP has been neglecting the privacy of the information exchanged by the agents during the computing of the solution, which is critical to many real-life problems. While agents are willing to cooperate with each other to produce an optimal solution, they are most often reluctant to reveal their private constraints and preferences to other, which hinders this cooperation. The goal of this project was to implement, test, and evaluate a secured version of DPOP, P-DPOP. P-DPOP provides strong agent privacy and topology privacy by randomization, constraints privacy and limited decision privacy by obfuscation. The algorithm was implemented in Java, as part of the open-source FRODO platform for DCOP.

2010

Chargement

On Vacuous and Non-Vacuous Generalization Bounds for Deep Neural Networks.

Konstantinos Pitas

Deep neural networks have been empirically successful in a variety of tasks, however their theoretical understanding is still poor. In particular, modern deep neural networks have many more parameters than training data. Thus, in principle they should overfit the training samples and exhibit poor generalization to the complete data distribution. Counter intuitively however, they manage to achieve both high training accuracy and high testing accuracy. One can prove generalization using a validation set, however this can be difficult when training samples are limited and at the same time we do not obtain any information about why deep neural networks generalize well. Another approach is to estimate the complexity of the deep neural network. The hypothesis is that if a network with high training accuracy has high complexity it will have memorized the data, while if it has low complexity it will have learned generalizable patterns. In the first part of this thesis we explore Spectral Complexity, a measure of complexity that depends on combinations of norms of the weight matrices of the deep neural network. For a dataset that is difficult to classify, with no underlying model and/or no recurring pattern, for example one where the labels have been chosen randomly, spectral complexity has a large value, reflecting that the network needs to memorize the labels, and will not generalize well. Putting back the real labels, the spectral complexity becomes lower reflecting that some structure is present and the network has learned patterns that might generalize to unseen data. Spectral complexity results in vacuous estimates of the generalization error (the difference between the training and testing accuracy), and we show that it can lead to counterintuitive results when comparing the generalization error of different architectures. In the second part of the thesis we explore non-vacuous estimates of the generalization error. In Chapter 2 we analyze the case of PAC-Bayes where a posterior distribution over the weights of a deep neural network is learned using stochastic variational inference, and the generalization error is the KL divergence between this posterior and a prior distribution. We find that a common approximation where the posterior is constrained to be Gaussian with diagonal covariance, known as the mean-field approximation, limits significantly any gains in bound tightness. We find that, if we choose the prior mean to be the random network initialization, the generalization error estimate tightens significantly. In Chapter 3 we explore an existing approach to learning the prior mean, in PAC-Bayes, from the training set. Specifically, we explore differential privacy, which ensures that the training samples contribute only a limited amount of information to the prior, making it distribution and not training set dependent. In this way the prior should generalize well to unseen data (as it hasn't memorized individual samples) and at the same time any posterior distribution that is close to it in terms of the KL divergence will also exhibit good generalization.

EPFL2020

Chargement

On Vacuous and Non-Vacuous Generalization Bounds for Deep Neural Networks.

Konstantinos Pitas

EPFL2020