Abstract: A method and data processing system for making a machine learning model more resistant to adversarial examples are provided. In the method, an input for a machine learning model is provided. A randomly generated mask is added to the input to produce a modified input. The modified input is provided to the machine learning model. The randomly generated mask negates the effect of a perturbation added to the input for causing the input to be an adversarial example. The method may be implemented using the data processing system.

Abstract: A method and data processing system for making a machine learning model more resistant to adversarial examples are provided. In the method, an input for a machine learning model is provided. A randomly generated mask is added to the input to produce a modified input. The modified input is provided to the machine learning model. The randomly generated mask negates the effect of a perturbation added to the input for causing the input to be an adversarial example. The method may be implemented using the data processing system.

Abstract: In an aspect, a method can include generating a cyclic redundancy check code for a binary data item, using a generator polynomial; and masking, using polynomial addition, the binary data item with a binary mask. The method can also include at least one of: storing, by a microcircuit, the masked binary data item in a memory of an electronic device; or transferring, by the microcircuit, the masked data item to another device. The cyclic redundancy check code for the binary data item can be generated from the masked binary data item to prevent discovery of the binary data item by a side-channel attack during the generating the cyclic redundancy check. The binary mask can be a multiple of a random number and the generator polynomial, such that respective cyclic redundancy check code of the masked data item and the binary data item have a same result.

Abstract: A cryptographic data processing method, implemented in an electronic device including a processor, the method including steps of providing a point of an elliptic curve in a Galois field, and a whole number, and of calculating a scalar product of the point by the number, the coordinates of the point and the number having a size greater than the size of words that may be processed directly by the processor, the scalar multiplication of the point by the number including steps of: storing scalar multiples of the point multiplied-by the number 2 raised to a power belonging to a series of whole numbers, setting a resulting point for each non-zero bit of the first number, adding the resulting point and one of the stored multiple points, and providing at the output of the processor the resulting point as result of the scalar product.

Abstract: The present invention relates to a method for performing an iterative calculation of exponentiation of a large datum, the method being implemented in an electronic device (DV1) and comprising calculations of squaring and multiplying large variables performed in parallel, by squaring (SB1) and multiplication (SM1) blocks, the method comprising steps of: while a temporary storage buffer memory is not full of unused squares, triggering a calculation by the squaring block for a bit of the exponent, when the squaring block is inactive, storing each square provided by the squaring block in the buffer memory, if the bit of the corresponding exponent is on 1, and while the buffer memory contains an unused square, triggering a calculation by the multiplication block concerning the unused square, when the multiplication block is inactive.

Abstract: A cryptographic data processing method, implemented in an electronic device including a processor, the method including steps of providing a point of an elliptic curve in a Galois field, and a whole number, and of calculating a scalar product of the point by the number, the coordinates of the point and the number having a size greater than the size of words that may be processed directly by the processor, the scalar multiplication of the point by the number including steps of: storing scalar multiples of the point multiplied-by the number 2 raised to a power belonging to a series of whole numbers, setting a resulting point for each non-zero bit of the first number, adding the resulting point and one of the stored multiple points, and providing at the output of the processor the resulting point as result of the scalar product.

Abstract: The present invention relates to a method for processing a binary data item, comprising a step of calculating a cyclic redundancy check code for the data item by means of a generator polynomial, wherein the step of calculating the cyclic redundancy check code comprises the steps of: masking the data item with a random binary mask that is a multiple of the generator polynomial, and generating the cyclic redundancy check code for the data item from the masked data item.

Abstract: The present invention relates to a method for performing an iterative calculation of exponentiation of a large datum, the method being implemented in an electronic device (DV1) and comprising calculations of squaring and multiplying large variables performed in parallel, by squaring (SB1) and multiplication (SM1) blocks, the method comprising steps of: while a temporary storage buffer memory is not full of unused squares, triggering a calculation by the squaring block for a bit of the exponent, when the squaring block is inactive, storing each square provided by the squaring block in the buffer memory, if the bit of the corresponding exponent is on 1, and while the buffer memory contains an unused square, triggering a calculation by the multiplication block concerning the unused square, when the multiplication block is inactive.

Abstract: An integrated circuit including a multiplication function configured to execute a multiplication operation of two binary words x and y including a plurality of basic multiplication steps of components xi of word x by components yj of word y is described. The multiplication function of the integrated circuit is configured to execute two successive multiplications by modifying, in a random or pseudo-random manner, an order in which the basic multiplication steps of components xi by components yj are executed.

Abstract: A process for testing an integrated circuit includes collecting a set of points of a physical property while the integrated circuit is executing a multiplication, dividing the set of points into a plurality subsets of lateral points, calculating an estimation of the value of the physical property for each subset, and applying to the subset of lateral points a step of horizontal transversal statistical processing by using the estimations of the value of the physical property, to verify a hypothesis about the variables manipulated by the integrated circuit.

Abstract: A method and a device protected against hidden channel attacks includes a calculation of the result of the exponentiation of a data m by an exponent d. The method and the device are configured to execute only multiplications of identical large variables, by breaking down any multiplication of different large variables x, y into a combination of multiplications of identical large variables.

Abstract: An integrated circuit including a multiplication function configured to execute a multiplication operation of two binary words x and y including a plurality of basic multiplication steps of components xi of word x by components yj of word y is described. The multiplication function of the integrated circuit is configured to execute two successive multiplications by modifying, in a random or pseudo-random manner, an order in which the basic multiplication steps of components xi by components yj are executed.

Abstract: A process for testing an integrated circuit includes collecting a set of points of a physical property while the integrated circuit is executing a multiplication, dividing the set of points into a plurality subsets of lateral points, calculating an estimation of the value of the physical property for each subset, and applying to the subset of lateral points a step of horizontal transversal statistical processing by using the estimations of the value of the physical property, to verify a hypothesis about the variables manipulated by the integrated circuit.