No abstract

The pervasiveness of camera technology in every-day life begets a modern reality in which images of individuals are routinely captured on a daily basis. Although this has enabled many benefits, it also infringes on personal privacy. To mitigate the loss of privacy, researchers have investigated methods of facial obfuscation in images. A promising direction has been the work in the k-same family of methods which employ the concept of k-anonymity from database privacy. However, there are a number of deficiencies of k-anonymity which carry over to the k-same methods, detracting from their usefulness in practice. In this paper, we first outline several of these deficiencies and discuss their implications in the context of facial obfuscation. We then develop the first framework to apply the formal privacy guarantee of differential privacy to facial obfuscation in generative machine learning models for images. Next, we discuss the theoretical improvements in the privacy guarantee which make this approach more appropriate for practical usage. Our approach provides a provable privacy guarantee which is not susceptible to the outlined deficiencies of k-same obfuscation and produces photo-realistic obfuscated output. Finally, while our approach provides a stronger privacy guarantee, we demonstrate through experimental comparisons that it can achieve comparable utility to k-same approaches in the context of preservation of demographic information in the images. The preservation of such information is of particular importance for enabling effective data mining on the obfuscated images.

No abstract

In this work, the problem of anomaly detection in imbalanced datasets, framed in the context of network intrusion detection is studied. A novel anomaly detection solution that takes both data‐level and algorithm‐level approaches into account to cope with the class‐imbalance problem is proposed. This solution integrates the auto‐learning ability of Reinforcement Learning with the oversampling ability of a Conditional Generative Adversarial Network (CGAN). To further investigate the potential of a CGAN, in imbalanced classification tasks, the effect of CGAN‐based oversampling on the following classifiers is examined: Naïve Bayes, Multilayer Perceptron, Random Forest and Logistic Regression. Through the experimental results, the authors demonstrate improved performance from the proposed approach, and from CGAN‐based oversampling in general, over other oversampling techniques such as Synthetic Minority Oversampling Technique and Adaptive Synthetic.

With large volumes of detailed health care data being collected, there is a high demand for the release of this data for research purposes. Hospitals and organizations are faced with conflicting interests of releasing this data and protecting the confidentiality of the individuals to whom the data pertains. Similarly, there is a conflict in the need to release precise geographic information for certain research applications and the requirement to censor or generalize the same information for the sake of confidentiality. Ultimately the challenge is to anonymize data in order to comply with government privacy policies while reducing the loss in geographic information as much as possible. In this paper, we present a novel geographic-based system for the anonymization of health care data. This system is broken up into major components for which different approaches may be supplied. We compare such approaches in order to make recommendations on which of them to select to best match user requirements. trade-off between the level of protection that can be achieved on a data set and the resultant utility of the data [13,14,17,34]. Although it is desirable to minimize the loss of any type of information in the data set, in some cases the preservation of geographic information may be of particular interest. Studies which involve the propagation of diseases across geographic areas require a high level of precision in the geographic information of the data set [23]. Any form of location-critical research such as spatial epidemiology requires high precision geographic information in order to be carried out [5,28,36]. However, the release of these precise geographic details greatly increases the risk of disclosure of confidential information due to higher levels of distinctness in the records of the data set. This risk creates a barrier in the disclosure of essential geographic information.In this paper, we present a novel and configurable system to achieve k-anonymity [34,35] on a data set through the use of geographic partitioning guided by the use of Voronoi diagrams [22]. This system, namedVoronoi-Based Aggregation System (VBAS), achieves anonymity in a data set through the generalization (coarsening of the level of precision) of geographic attributes and the suppression of records. By aggregating regions, we avoid the need for the suppression of small regions, which can lead to heavily censored data sets [10,20], while maintaining a higher degree of geographic precision than other methods (such as cropping [7,21]).Since any loss in geographic information has negative effects on the ability to effectively analyze a data set, we postulate it is desirable to preserve as much geographic information as possible [27]. VBAS addresses this problem by aggregating small regions of fine granularity into larger regions that satisfy criteria for achieving a sufficient level of anonymity while reducing the loss of geographic information. In order to evaluate the quality of the resultant aggregation, we employ measures of suppression an...