IRDETO B.V.

Disclosed is a method for identifying whether a classification system is configured to use a specific machine-learning classification model. The method involves using the classification system to generate, for each test sample in a predetermined test set, a corresponding classification result, and identifying either (i) that the classification system is using the specific machine-learning classification model or (ii) that the classification system is not using the specific machine-learning classification model. The test set is associated with the specific machine-learning classification model and, for each test sample in the test set, there is a corresponding small modification for that test sample that causes a change in the classification result produced for that test sample using the specific machine-learning classification model.

There is described a method, at a server system, of providing a mobile network operator (MNO) profile to a client device. The client device has a SIM software application stored thereon so as to provide the client device with a secured software implementation of SIM card functionality. The method comprises: (a) based on a unique identifier of the client device, identifying a unique key, K SIM , of the SIM software application stored on the client device; (b) based on an MNO associated with the client device, identifying an unused MNO profile associated with the MNO; (c) encrypting the identified MNO profile so as to provide an encrypted MNO profile, wherein the encrypting comprises encrypting at least part of the identified MNO profile using K SIM ; (d) generating an MNO profile download message comprising the encrypted MNO profile and the unique identifier of the client device; and (e) broadcasting the MNO profile download message over a broadcast network so as to enable the client device to access the MNO profile download message. There is also described a related method at a client device, as well as related computer programs and computer-readable media.

A method and system for improving the efficiency of locating a pool operator in a Plug and Charge ecosystem. The method includes receiving an identifier associated with a user, using a Domain Name System (DNS) as a reference engine for the received identifier, generating a fully qualified domain name using the received identifier, and directing the fully qualified domain name to an internet address of a corresponding pool operator. The system includes one or more processors, a DNS, a fully qualified domain name generator, and a directing module. A non-transitory computer-readable medium storing instructions for performing the method is also disclosed.

Disclosed implementations include a method of detecting attacks on Machine Learning (ML) models by applying the concept of anomaly detection based on the internal state of the model being protected. Instead of looking at the input or output data directly, disclosed implementation look at the internal state of the hidden layers of a neural network of the model after processing of data. By examining how different layers within a neural network model are behaving an inference can be made as to whether the data that produced the observed state is anomalous (and thus possibly part of an attack on the model).

A method and system for managing multiple Plug and Charge mobility service provider charging contracts outside a vehicle by installing a single contract certificate that can map to one or many mobility service provider charging contracts. After the authentication of the contract certificate, the mapping of the contract certificate to a charging contract is accomplished.

A method for a testing system to perform fuzzy testing of a software system, wherein the software system comprises a plurality of callable units and is arranged to receive input for the software system to process, the method comprising: determining, for each callable unit of the plurality of callable units, based on one or more security vulnerability metrics, a target number of times that callable unit is to be tested; initializing a ranked plurality of queues, each queue for storing one or more seeds, said initializing comprising storing one or more initial seeds in a corresponding queue of the ranked plurality of queues; performing a sequence of tests, wherein performing each test comprises: obtaining a seed from the highest ranked non-empty queue; performing a mutation process on the obtained seed to generate a test seed, wherein the mutation process is configured, at least in part, by mutation guidance information; providing the test seed as input to the software system for the software system to process; and evaluating the processing of the test seed by the software system to generate a result for the test; wherein each queue in the ranked plurality of queues has an associated seed addition criterion and wherein performing each test comprises either (a) adding the test seed to the highest ranked queue in the ranked plurality of queues for which the test seed meets the seed addition criterion associated with that queue; or (b) discarding the test seed if the test seed does not meet the seed addition criterion associated with any of the queues in the ranked plurality of queues; wherein the seed addition criteria are configured so that, if processing of a first test seed by the software system involves execution of, or an execution path approaching, a callable unit of interest and if processing of a second test seed by the software system does not involve execution of, or an execution path approaching, a callable unit of interest, then the queue to which the first test seed is added is of higher rank than the queue to which the second test seed is added, wherein a callable unit is a callable unit of interest if the current number of tests that have resulted in execution of that callable unit is less than the target number of times that callable unit is to be tested.

A method of performing a cryptographic process in a secured manner, wherein the cryptographic process generates output data based on input data, the generating of the output data involving generating a value y based on an amount of data x, the value y representing a combination, according to a linear transformation L, of respective outputs from a plurality of S-boxes S n (n=0, . . . , N−1) for integer N>1, wherein each S-box S n (n=0, . . . , N−1) implements a respective function H n that is either (a) the composition of a respective first function F n and a respective linear or affine second function G n so that H n =G n ∘F n , or (b) the composition of a respective first function F n , a respective linear or affine second function G n and a respective third function W n so that H n =G n ∘F n ∘W n , wherein the method comprises: performing a first processing stage and a second processing stage to generate the value y based on the amount of data x, wherein: the first processing stage uses a plurality of first lookup tables to generate respective outputs, each output being based on at least part of the amount of data x, wherein, for each S-box S n (n=0, . . . , N−1), the respective first function F n is implemented by a corresponding first lookup table; and the second processing stage combines outputs from a plurality of second lookup tables to generate the value y, wherein the input to each second lookup table is formed from the output of a plurality of the first lookup tables, and wherein the set of second lookup tables is based on the second functions G n (n=0, . . . , N−1) and the linear transformation L.

Disclosed implementations include a method, apparatus and computer media for learning an optimal graph in the form of a tree topology defining a sequence that can be used by a learning network for image recognition. Image data representing the image of an object is received and N landmarks are detected on the image using a deep regression algorithm, wherein N is an integer. A weighted, fully connected, graph is constructed from the landmarks by assigning initial weights for the landmarks randomly. An optimized tree structure is determined based on the initial weights. A sequence is generated by traversing nodes of the tree structure and a series of embeddings representing the object image are generated based on the sequence. The embeddings can be processed by a neural network to generate an image recognition signal based on the embeddings.

A method and system for improving the efficiency of locating a pool operator in a Plug and Charge ecosystem. The method includes receiving an identifier associated with a user, using a Domain Name System (DNS) as a reference engine for the received identifier, generating a fully qualified domain name using the received identifier, and directing the fully qualified domain name to an internet address of a corresponding pool operator. The system includes one or more processors, a DNS, a fully qualified domain name generator, and a directing module. A non-transitory computer-readable medium storing instructions for performing the method is also disclosed.

There is provided systems and methods for operating a keyless vehicle system of a vehicle. One such method comprises, receiving, at a Bluetooth module of the keyless vehicle system a connection request from a mobile device, wherein the Bluetooth module of the keyless vehicle system is bonded as a human interface device (HID) to the mobile device. In response to the mobile device being verified by the Bluetooth module, a Bluetooth connection between the mobile device and the Bluetooth module is established. In response to the vehicle detecting physical interaction, the presence of the mobile device determined. Operation of the keyless vehicle system is allowed based on the presence of the mobile device.