Abstract: The invention relates to a method for pooling and analyzing datasets from data providers. The method includes, at the level of each data provider, for each dataset, a first encryption of each variable of the dataset with a first encryption key common to all data providers, and a second encryption of at least each variable of the dataset with a second encryption key specific to the each data provider and shared with an intermediary. The method also includes, at the level of the intermediary, for each encrypted dataset, decryption of the variables with the second encryption key, and a third encryption of each confidential variable with a third encryption key. The method also includes transmission of the datasets encrypted with the third encryption key to an analysis environment for analysis. It also relates to a computer program and a system for implementing such a method.

Abstract: The invention relates to a federated learning method including storing, in a central node and in N local nodes, an intermediate model obtained by training an object detection model based on a central training dataset; for each local node, obtaining a local model by training the intermediate model based on a private training dataset; for each image of a public dataset, and for each local model, determining a prediction vector and an attention map; for each image of the public dataset, generating an aggregated set of outputs from the prediction vectors, and an aggregated attention map from the attention maps; and obtaining a central model by modifying parameters of the intermediate model stored in the central node to minimize a difference between the aggregated set of outputs and an output of the intermediate model based on the public dataset as input.

Abstract: method for processing a computer job to be executed on a hybrid computer configured with processing resources comprising one or more central processing unit, CPU, processing resources and at least one quantum processing unit, QPU, processing resources, the method comprising, wherein the computer job comprises a quantum computation part to be executed on the at least one QPU processing resources: Processing the computer job by a first scheduler of the hybrid computer, which comprises: Upon receiving a request for quantum processing resources for execution of the computer job, offloading the scheduling of a portion of the computer job comprising the quantum computing part on processing resources of the hybrid computer to a second scheduler of the hybrid computer; and receiving, by the first scheduler, status data regarding execution of the portion of the computer job comprising the quantum computing part.

Abstract: The invention relates to a method for determining calibration data, for a goniometry apparatus comprising an antenna array, from several sets of calibration data. The sets are associated with a measured angular position and include an amplitude datum and a phase datum measured by each antenna. The method includes calculating an estimated calibration data set and normalizing the data set measured by providing as many normalized data sets as there are antennas for each measured position. The method includes calculating a candidate data set by interpolating the measured data sets at the measured positions and previously normalized with respect to the phase measured by the antenna, and selecting the candidate data set whose phase reference has the highest energy. The invention relates to a computer program and an apparatus implementing the method, a calibration table obtained by the method and a goniometry apparatus calibrated with the method.

Abstract: A method for detecting outliers in processes running in a group of machines. A clustering stage, carried out at a first frequency, including fetching a list of software contained in all machines, calculating tf-idf value for each installed software and for each machine, performing clustering of the machines by applying a clustering algorithm and using a Jaccardian weighted distance method between machines based on the tf-idf values. A preliminary outliers detection stage, carried out at a second frequency greater than the first frequency, including fetching information of processes running in the machines, for each cluster calculating tf-idf values for each process, wherein if a tf-idf value is greater than a first predetermined threshold, the process is considered as outlier, for all clusters calculating itf-idf value for each process considered as outlier, and if a itf-idf value is lower than a second predetermined threshold, the process is confirmed as outlier.

Abstract: Systems and methods are described for license management. A system for license management includes a smart contract that records a licensing agreement between a first party and a second party on usage of a network function. The smart contract is persisted in a trade blockchain. The system also includes an e-licensing manager agent. The e-licensing manager agent retrieves the licensing agreement from the smart contract. The e-licensing manager agent monitors a usage of the network function by the second party in real-time based on the licensing agreement, record actions of the usage of the network function, and create a transaction record based on the recorded actions in an operation blockchain.

Abstract: The invention relates to a system and method for managing the data streams for unified governance of a plurality of intensive computing solutions (70) accessible to a user client (2) from an aggregated interface (10), said intensive computing solutions including at least two solutions selected among: a high performance computing server (71), a server dedicated to supervised or unsupervised learning (72) and a server dedicated to quantum computing (73); said method being implemented at least partly by said computer system including: at least one database (25a) configured to store execution data for intensive computing operations; said method including the steps of storing (300) execution data for the intensive computing solutions (70) and transmitting (400) execution data for the intensive computing solutions to the aggregated interface (10).

Abstract: A computer-implemented method for generating a program to be executed using a quantum computer for producing as output a target quantum state based on an initial quantum state used as input is proposed, which comprises: determining an approximated quantum state which is an approximation of the target quantum state; determining a quantum circuit which, based on the initial quantum state received used as input, produces as output an output quantum state that corresponds to the approximated quantum state; and generating the program based on the determined circuit.

Abstract: The invention relates to a software-defined radio (200) comprising: at least one processor (104) producing a digital signal, a digital-to-analog converter, DAC (106), to convert said digital signal into an analog signal, a physical interface (110) between the processor (104) and the DAC, and a radio-frequency power amplifier (108) to amplify the power of said analog signal; said software-defined radio (200) further comprising, arranged in said physical interface (110), a unit (202) for protecting said power amplifier (108) depending on: the average power, over a predefined reference period, of the digital signal received by said physical interface (110), a predefined power threshold not to be exceeded during said period. It also relates to an electronic device comprising such a digital radio and a method of protecting a software-defined radio.

Abstract: The invention relates to a method for determining an environmental impact of a product, including obtaining a list of components of the product, for each component, requesting a list of materials from a life cycle inventory database, generating a request comprising the list of materials that is obtained for selecting a most suitable material for the product, sending the request to a language model module, receiving a name of the most suitable material for the product, for each name of each material of each component, requesting the life cycle inventory database to send at least one environmental impact value, receiving the at least one environmental impact value for each name of each material of each component, and calculating a total environmental impact of the product using the at least one environmental impact value that is received for each component and their respective mass in the product.

Abstract: The invention relates to an edge-based log collecting device for collecting logs from several log sources located in a remote network, called edge-network, and sending the logs to a cloud-based system distant from the edge-network. The device is located in the edge-network, and the device includes several processing nodes for processing logs received from the log sources and sending them to said the system. The device is configured to elect, according to a predetermined election algorithm, one of the processing nodes as a master node configured for receiving the logs from all log sources of the local network, and sharing the logs with the other processing nodes. The invention further relates to a computer program and a device configured to carry out such a method.

Abstract: The invention relates to a method for allowing at least one virtual reality device to access a data platform of a three-dimensional content system, said method comprising in particular the steps of receiving a cryptographic footprint on a control communication link (CCL), receiving a cryptographic footprint of an application stored on the at least one virtual reality device sent by said at least one virtual reality device on a access communication link (ACL), comparing both received cryptographic footprints, sending a network address of a selected application environment and authentication data to the virtual reality device on the access communication link (ACL), creating, by the at least one virtual reality device, a secured communication between said at least one virtual reality device and the selected application environment.

Abstract: A system that detects malware by analyzing message logs to identify message patterns that are periodic with similar-sized messages. These patterns may indicate malware since malware often sends beacon messages to a command-and-control system that are often periodic and of relatively similar length. The system may group message logs by the combination of source and destination and analyze each group for patterns of periodicity and message length uniformity. Entropy may be used to measure the uniformity of message lengths and message intervals, with low (or zero) entropy suggesting malware. Message intervals that repeat after several messages may be detected by testing subsequence sums for uniformity at different possible periods. Additional factors may be used to assess the risk, such as the duration of communication, and threat intelligence on the source or destination. The system may perform automated actions to eliminate or mitigate detected risks, such as blocking further communication.

Abstract: The invention relates to a method implemented by a computer for selecting a storage configuration defining at least one storage medium to use during the execution of an application to be executed by the high-performance computer. The method including obtaining input/output data of the application, decomposing the obtained input/output data into storage phases and computational phases, obtaining expected storage performances comprising providing the storage phases to a trained performance regression model, simulating the execution of the first application for each storage configuration of a plurality of storage configurations based on the plurality of storage mediums, wherein the simulation includes providing a performance score of each storage configuration of the plurality of storage configurations, and selecting a storage configuration having a highest performance score for the execution of the application.

Abstract: A method for detecting malicious connections from remote users into a computer network through Remote Desktop protocol via a computer having access to login logs of users. The method includes defining aspects, each divided into bins comprising a day of week aspect comprising n1 bins, a time of day aspect comprising n2 bins, a number of logins in a day aspect comprising n3 bins. The method includes defining a model based on the aspects and providing a score of log for each user; defining a baseline of log; applying the model on each user log to determine a production score of log and comparing the production score of log with respect to the baseline. The model includes calculating a probability density for each bin for each user, determining a weight for each aspect and calculating the score of log from the probability density weighted by the weight for each user.

Abstract: Method for determining an approximated final state having a fidelity above a bound, comprising: receiving an initial state; receiving a quantum circuit comprising gates; defining a lower bound of the fidelity; iterating over the gates: applying a current gate to the initial state; if the current gate is a two-qubit gate, factorizing a portion of the updated state by SVD into a product of a unitary matrix, a diagonal matrix, and a unitary matrix; truncating the bond dimension of the diagonal matrix to a target bond dimension such that a product of a truncation fidelity of the truncated matrix, truncation fidelities of previously truncated matrices and future target truncation fidelities is greater than the bound; in a next iteration, using the updated state as initial state; defining an approximated final state equal to the updated state of a last iteration.

Abstract: The invention relates to an electronic board (200) comprising: a support (102), at least one electronic component (104) mounted on said support (102), a protective layer (110), mounted on said support (102), covering said electronic component (104) at a non-zero distance from said electronic component (104), and a heat dissipation device for said electronic component (104) comprising a metal thermal connecting part (202) screwed into said protective layer (110), passing through said protective layer (110) and coming into contact with said electronic component (104), so as to create a thermal connection between said electronic component (104) and said protective layer (110). It also relates to an electronic apparatus comprising such an electronic board.

Abstract: The invention relates to a method for replacing liquid in a liquid cooling circuit in a supercomputer rack, including a main pump and a reserve pump, using a maintenance unit. The method includes disconnecting the reserve pump, connecting the upstream reservoir to the reserve pump, connecting the downstream reservoir, stopping the main pump, activating the reserve pump, stopping the reserve pump once the upstream reservoir is below a predetermined low threshold, disconnecting the upstream reservoir from the reserve pump, disconnecting the reserve pump from the downstream reservoir, connecting the inlet supply pipe to the reserve pump in order to return said reserve pump to its initial configuration and reactivating the main pump.

Abstract: Method for determining a lower bound of a fidelity of an approximated final state, comprising: receiving an initial state in a matrix product representation; receiving a quantum circuit comprising gates; iterating over the gates: applying a current gate to the initial state; if the current gate is a two-qubit gate, factorizing a portion of the updated state by SVD into a product of a unitary matrix, a diagonal matrix, and a unitary matrix; if a bond dimension of the diagonal matrix exceeds a threshold: truncating the diagonal matrix such that the bond dimension does not exceed said threshold, and determining a truncation fidelity; in a next iteration, using the updated state as the initial state; determining a lower bound of the fidelity of the approximated final state as a product of the truncation fidelities.

Abstract: A method for manufacturing a series of parts using a manufacturing machine of a production line and a supervision device configured for controlling in real time the manufacturing machine taking into account quality data, carbon footprint and cost. The method includes receiving real time data from the production line, running a first prediction algorithm to predict quality data on the series of parts in real time, running a second prediction algorithm to predict a carbon footprint of the series of parts in real time, running a third prediction algorithm to predict the cost of the series of parts in real time, determining a set of scenarios based on the predicted carbon footprint and the predicted cost, selecting at least one scenario in the set of scenarios based on the predicted carbon footprint and/or the predicted cost, manufacturing the series of parts according to the selected scenario.