Here Global B.v.

Raghavendran Balu

Arash Ostadzadeh

Matei Stroila

Bo Xu

Nirav Ajmeri

Stefano Bennati

Elena Vidyakina

Elena Mumford

Zack Zhu

Aleksandra Kovacevic

Nicholas Dronen

Kai Pöthkow

Shubhabrata Roy

Amritpal Singh Gill

A method, apparatus, and computer program product are provided for providing improved neural network implementations using a repeated convolution-based attention module. Example embodiments implement a repeated convolution-based attention module that utilizes multiple iterations of a repeated convolutional application layer and subsequent augmentations to generate an attention module output. Example methods may include augmenting an attention input data object based on a previous iteration convolutional output to produce a current iteration input parameter, inputting the input parameter to a repeated convolutional application layer to generate a current iteration input parameter, repeating for multiple iterations, and augmenting the attention input data object based on the final convolutional output to produce an attention module output. Other methods may include an initial convolutional application layer, and/or apply and augment the output of the initial convolutional application layer, and include convolutional application layer(s) having at least two sub-layers.

Methods, apparatuses, and computer program products using a repeated convolution-based attention module for improved neural network implementations

An apparatus for providing anonymity in geographic data for probe devices in a geographic region for a location-based service includes at least a database, a clustering calculator and an anonymity controller. The database stores trajectory data based on sequences of sensor measurements of the probe devices. The clustering calculator clusters the trajectory data, according to a first iteration threshold, into clusters each defined by a cluster point and compares distance for a first cluster from the clusters to cluster points of other clusters of the clusters. The clustering calculator selects a second cluster from the clusters based on the comparison of distances and merges the first cluster and the second cluster into a merged cluster. The anonymity controller modifies the trajectory data to provide a predetermined level of anonymity to locations from the trajectory data in response to the merged cluster.

Clustering for K-anonymity in location trajectory data

Systems and methods for reconstructing a trajectory from anonymized data are provided. In some aspects, a method includes receiving anonymized data corresponding to a trajectory of a user or object, and assembling, based on the anonymized data, a state-space model. The method also includes executing a prediction algorithm, based on the state-space model, to generate predicted data from the anonymized data, and reconstructing the trajectory of the user or object using the predicted data. The method further includes generating a report indicative of the trajectory.

Systems and methods for reconstructing a trajectory from anonymized data

A method, apparatus and computer program product are provided to determine a semantic privacy index that quantifies the risk associated with re-identification of a trajectory following anonymization of the trajectory. In the context of a method, information regarding a trajectory is received. After the trajectory has been map matched to a portion of a road network, the method associates contextual information comprising one or more map features with the trajectory. The method also provides the information regarding the trajectory and the contextual information comprising the one or more map features to a risk estimation model in order to generate the semantic privacy index.

Method, apparatus and computer program product for determining a semantic privacy index

An apparatus for the quantification of privacy risk in geographic data for probe devices in a geographic region includes a database, a sub-trajectory module, a candidate list calculator, and a provision module. The database is configured to store trajectory data based on sequences of sensor measurements. The sub-trajectory module is configured to receive trajectory data points based on sequences of sensor measurements of the probe devices collected in the geographic region and determine sub-trajectories from changes in trajectory identifiers. The candidate list calculator is configured to concatenate at least two of the sub-trajectories based on at least one concatenation parameter. The provision module is configured to calculate a reconstruction rate in the trajectory data and provide the reconstruction rate for a quantification of privacy risk to an external device.

Quantification of privacy risk in location trajectories

An apparatus for providing anonymity in geographic data for probes using a variable mix zone in a geographic region includes a geographic database, a mix zone generator, and a pseudonym generator. The geographic database is configured to store one or more map features in the geographic region. The mix zone generator is configured to define a mix zone boundary having a size based on the one or more map features and monitor probe data for at least one probe. The probe data including a location within a predetermined distance of the mix zone boundary. The pseudonym generator is configured to select a pseudonym for the at least one probe in response to the location of the probe data and apply the pseudonym for the at least one probe in the geographic data when the at least one probe exits the mix zone.

Dynamic mix zones

An apparatus for providing anonymity in geographic data for probe devices in a geographic region for a location-based service includes at least a database, a map matcher, and an aggregation calculator. The database is configured to store trajectory data based on sequences of sensor measurements of the probe devices. The map matcher is configured to match a map, including cells, with the trajectory data, and each of the cells includes edges and a face. The aggregation calculator configured to calculate a number of trajectories in the trajectory data that passes through each of the edges and compare the number of trajectories in the trajectory data for at least one of the edges to a threshold. The geographic data is modified in response to the comparison of the number of trajectories in the trajectory data for the at least one of the edges.

Map matched aggregation for K-anonymity in trajectory data

An apparatus for providing anonymity in geographic data for probe devices in a geographic region for a location-based service includes at least a database and a triangulation calculator. The database is configured to store trajectory data based on sequences of sensor measurements of the probe devices. The triangulation calculator is configured to calculate triangles from the trajectory data. Each of the triangles is defined by vertices. The triangulation calculator is configured to calculate a similarity score for the vertices based on trajectory data associated with adjacent vertices. At least one vertex is modified in response to the similarity score.

Growing community of inventors

Eindhoven, Netherlands

Raghavendran Balu