Claudio Silvestri

Preventing velocity-based linkage attacks in location-aware applications

Mobile devices with positioning capabilities allow users to participate in novel and exciting location-based applications. For instance, users may track the whereabouts of their acquaintances in location-aware social networking applications, e.g., GoogleLatitude. Furthermore, users can request information about landmarks in their proximity. Such scenarios require users to report their coordinates to other parties, which may not be fully trusted. Reporting precise locations may result in serious privacy violations, such as disclosure of lifestyle details, sexual orientation, etc. A typical approach to preserve location privacy is to generate a cloaking region (CR) that encloses the user position. However, if locations are continuously reported, an attacker can correlate CRs from multiple timestamps to accurately pinpoint the user position within a CR. In this work, we protect against linkage attacks that infer exact locations based on prior knowledge about maximum user velocity. Assume user u who reports two consecutive cloaked regions A and B. We consider two distinct protection scenarios: in the first case, the attacker does not have information about the sensitive locations on the map, and the objective is to ensure that u can reach some point in B from any point in A. In the second case, the attacker knows the placement of sensitive locations, and the objective is to ensure that u can reach any point in B from any point in A. We propose spatial and temporal cloaking transformations to preserve user privacy, and we show experimentally that privacy can be achieved without significant quality of service deterioration.

Fine-Grained Cloaking of Sensitive Positions in Location-Sharing Applications

Geosocial networking applications magnify the concern for location privacy because a users position can be disclosed to diverse untrusted parties. The Privacy Preserving Obfuscation Environment (Probe) framework supports semantic-location cloaking to protect this information.

A new algorithm for gap constrained sequence mining

The sequence mining problem consists in finding frequent sequential patterns in a database of time-stamped events. Several application domains require limiting the maximum temporal gap between events occurring in the input sequences. However pushing down such constraint is critical for most sequence mining algorithms.In this paper we describe CCSM (Cache-based Constrained Sequence Miner), a new level-wise algorithm that overcomes the troubles usually related to this kind of constraints. CCSM adopts an innovative approach based on k-way intersections of idlists to compute the support of candidate sequences. Our k-way intersection method is enhanced by the use of an effective cache that stores intermediate idlists for future reuse. The reuse of intermediate results entails a surprising reduction in the actual number of join operations performed on idlists.CCSM has been experimentally compared with cSPADE, a state of the art algorithm, on several synthetically generated datasets, obtaining better or similar results in most cases.

Spatio-temporal aggregations in trajectory data warehouses

In this paper we investigate some issues related to the design of a simple Data Warehouse (DW), storing several aggregate measures about trajectories of moving objects. First we discuss the loading phase of our DW which has to deal with overwhelming streams of trajectory observations, possibly produced at different rates, and arriving in an unpredictable and unbounded way. Then, we focus on the measure presence, the most complex measure stored in our DW. Such a measure returns the number of trajectories that lie in a spatial region during a given temporal interval. We devise a novel way to compute an approximate, but very accurate, presence aggregate function, which algebraically combines a bounded amount of measures stored in the base cells of the data cube.

Privacy-Preserving Sharing of Sensitive Semantic Locations under Road-Network Constraints

This paper presents a privacy-preserving framework for the protection of sensitive positions in real time trajectories. We assume a scenario in which the sensitivity of users positions is space-varying, and so depends on the spatial context, while the users movement is confined to road networks and places. Typical users are the non-anonymous members of a geo-social network who agree to share their exact position whenever such position does not fall within a sensitive place, e.g. a hospital. Suspending location sharing while the user is inside a sensitive place is not an appropriate solution because the users stopovers can be easily inferred from the users trace. In this paper we present an extension of the semantic location cloaking model [1] originally developed for the cloaking of non-correlated positions in an unconstrained space. We investigate different algorithms for the generation of cloaked regions over the graph representing the urban setting. We also integrate methods to prevent velocity-based linkage attacks. Finally we evaluate experimentally the algorithms using a real data set.

Approximate Aggregations in Trajectory Data Warehouses

In this paper we discuss how data warehousing technology can be used to store aggregate information about trajectories and perform OLAP operations over them. To this end, we define a data cube with spatial and temporal dimensions, discretized according to a regular grid. We investigate in depth some issues related to the computation of a holistic aggregate function, i.e, the presence, which returns the number of distinct trajectories occurring in a given spatio-temporal area. In particular, we introduce a novel way to compute an approximate, but nevertheless very accurate, presence aggregate function, which uses only a bounded amount of measures stored in the base cells of our cuboid. We also concentrate on the loading phase of our data warehouse, which has to deal with an unbounded stream of trajectory observations. We suggest how the complexity of this phase can be reduced, and we analyse the errors that this procedure induces at the level of the sub-aggregates stored in the base cells. These errors and the accuracy of our approximate aggregate functions are carefully evaluated by means of tests performed on synthetic trajectory datasets.

A general framework for trajectory data warehousing and visual OLAP

In this paper we present a formal framework for modelling a trajectory data warehouse (TDW), namely a data warehouse aimed at storing aggregate information on trajectories of moving objects, which also offers visual OLAP operations for data analysis. The data warehouse model includes both temporal and spatial dimensions, and it is flexible and general enough to deal with objects that are either completely free or constrained in their movements (e.g., they move along a road network). In particular, the spatial dimension and the associated concept hierarchy reflect the structure of the environment in which the objects travel. Moreover, we cope with some issues related to the efficient computation of aggregate measures, as needed for implementing roll-up operations. The TDW and its visual interface allow one to investigate the behaviour of objects inside a given area as well as the movements of objects between areas in the same neighbourhood. A user can easily navigate the aggregate measures obtained from OLAP queries at different granularities, and get overall views in time and in space of the measures, as well as a focused view on specific measures, spatial areas, or temporal intervals. We discuss two application scenarios of our TDW, namely road traffic and vessel movement analysis, for which we built prototype systems. They mainly differ in the kind of information available for the moving objects under observation and their movement constraints.

T-Warehouse: Visual OLAP analysis on trajectory data

Technological advances in sensing technologies and wireless telecommunication devices enable novel research fields related to the management of trajectory data. As it usually happens in the data management world, the challenge after storing the data is the implementation of appropriate analytics for extracting useful knowledge. However, traditional data warehousing systems and techniques were not designed for analyzing trajectory data. Thus, in this work, we demonstrate a framework that transforms the traditional data cube model into a trajectory warehouse. As a proof-of-concept, we implemented T-WAREHOUSE, a system that incorporates all the required steps for Visual Trajectory Data Warehousing, from trajectory reconstruction and ETL processing to Visual OLAP analysis on mobility data.

Protecting location privacy against spatial inferences: the PROBE approach

The widespread adoption of location-based services (LBS) raises increasing concerns for the protection of personal location information. A common strategy, referred to as obfuscation, to protect location privacy is based on forwarding the LSB provider a coarse user location instead of the actual user location. Conventional approaches, based on such technique, are however based only on geometric methods and therefore are unable to assure privacy when the adversary is aware of the geographical context. This paper provides a comprehensive solution to this problem. Our solution presents a novel approach that obfuscates the user location by taking into account the geographical context and users privacy preferences. We define several theoretical notions underlying our approach. We then propose a strategy for generating obfuscated spaces and an efficient algorithm which implements such a strategy. The paper includes several experimental results assessing performance, storage requirements and accuracy for the approach. The paper also discusses the system architecture and shows that the approach can be deployed also for clients running on small devices.

gpuDCI: Exploiting GPUs in Frequent Itemset Mining

Frequent item set mining (FIM) algorithms extract subsets of items that occurs frequently in a collection of sets. FIM is a key analysis in several data mining applications, and the FIM tools are among the most computationally intensive data mining ones. In this work we present a many-core parallel version of a state-of-the-art FIM algorithm, DCI, whose sequential version resulted, for most of the tested datasets, better than FP-Growth, one of the most efficient algorithms for FIM. We propose a couple of parallelization strategies for Graphics Processing Units (GPU) suitable for different resource availability, and we present the results of several experiments conducted on real-world and synthetic datasets.