Abstract: A computer platform includes an artificial neural network (ANN) as well as a classifier. The ANN is configured, after a learning phase, to transform an input data vector into a discriminating feature vector having a smaller dimension. A user then generates, from a plurality of reference data vectors, the same plurality of reference feature vectors, which are encrypted in an encryption module using the public key of a homomorphic cryptosystem and stored in a reference database of the platform. When the user requests the classification of an input data vector, the ANN, or a copy thereof, provides the classifier with a corresponding discriminating feature vector (y). Distances from the vector to the different reference feature vectors are calculated in the homomorphic domain and the index of the reference feature vector closest to y, i.e. the identifier i0 of the class to which it belongs, is returned to the user.

Abstract: A method for controlling access of a user equipped with a terminal to a physical or logical resource, the method involving a secure cryptographic device forming a token corresponding to an access criterion, the access token being intended to generate a keystream masking a biometric reference of the user obtained by a biometric reader of the terminal. The biometric reference thus masked is encrypted by fully homomorphic encryption and stored in a database hosted by a remote server. An access control operator obtains a biometric characteristic of the user, homomorphically encrypts it and transmits it to the remote server. This server compares the first and second biometric models in the homomorphic domain and supplies the homomorphically-encrypted result of the comparison to the access control operator. The latter grants or denies access to the user according to the result of the comparison, after having decrypted it.

Abstract: A method for confidentially querying the presence of a record in a database hosted by a server, the records being stored in the database in the form of digital footprints obtained by hashing a record by a public hash function. The footprints are masked by a stream cipher using a symmetric key of a first user. The first user may grant a second user authorisation to query the database by transmitting the inverse masks of various rows, encrypted by the public key of an additive homomorphic cryptosystem of the second user. The rows of the database are unmasked in the homomorphic domain and the second user transmits an encrypted request to query the base according to a PIR protocol. The second user can decrypt the response from the server using the private key of their homomorphic cryptosystem and determine whether the footprint sought is present in the response thus decrypted.

Abstract: The present invention concerns a method for secure data classification by a computer platform. A client sends to the platform data to be classified in encrypted form using a first symmetric key. Similarly, a supplier sends to the platform parameters of a classification model in encrypted form using a second symmetric key. The invention uses a homomorphic cryptosystem defined by a public key and a private key. The platform performs a first transcryption step by deciphering the data to be classified in the homomorphic domain and a second transcryption step by deciphering the model parameters in the homomorphic domain. The classification function is then evaluated in the homomorphic domain for providing a classification result encrypted by said public key.

Abstract: A method for federative learning of an artificial neural network model on a plurality of sets of training data. The learning method used a plurality of data suppliers each having a distinct set of training data and a symmetric key, and an aggregation platform aggregating partial models trained on a sub-plurality of these sets, in each iteration. In each iteration, the platform selects a sub-plurality of data suppliers and supplies the parameters of the model to them, in homomorphically encrypted form. Each training data supplier decrypts these parameters, trains the model on his own data set and returns the parameters of the partial model thus obtained, encrypted by his symmetric key, to the platform. The aggregation platform then transcrypts these partial models in the homomorphic domain, and combines them in the homomorphic domain to obtain a global model.

Abstract: A computer platform includes an artificial neural network (ANN) as well as a classifier. The ANN is configured, after a learning phase, to transform an input data vector into a discriminating feature vector having a smaller dimension. A user then generates, from a plurality of reference data vectors, the same plurality of reference feature vectors, which are encrypted in an encryption module using the public key of a homomorphic cryptosystem and stored in a reference database of the platform. When the user requests the classification of an input data vector, the ANN, or a copy thereof, provides the classifier with a corresponding discriminating feature vector (y). Distances from the vector to the different reference feature vectors are calculated in the homomorphic domain and the index of the reference feature vector closest to y, i.e. the identifier i0 of the class to which it belongs, is returned to the user.

Abstract: The present invention concerns a method for processing system logs of a computer system. A system log generator (LG) transmits these system logs to a system log analyser (SIEM) after they have been encrypted by means of a symmetric encryption key and sends the symmetric encryption key in parallel with a homomorphic cryptosystem public key. The system log analyser carries out a transcryption of these logs then a processing thereof in the homomorphic domain. The result of the processing in the homomorphic domain is then transmitted to a security centre (SOC) or even directly to the system log generator to be decrypted there. The security centre can establish a security report or propose a countermeasure before sending it, in form encrypted by the symmetric key, to the system log generator.

Abstract: A method for confidentially processing data stored on a platform is described. A transcryption module, receiving a request for access to a data block, transmits the access request to an access request prediction module which returns a list of data blocks. The transcryption module determines whether a corresponding key stream block is present in the cache. The homomorphic calculation module calculates homomorphic ciphertexts of key stream blocks corresponding to the data blocks of the list. The transcryption module transcrypts the data blocks by adding them with the homomorphic ciphertexts of the corresponding key stream blocks. The transcryption module transmits the data blocks thus transcrypted to the request management module.

Abstract: A circuit for generating twiddle factors for an NTT processor. The circuit includes a cache management manager, a modular multipliers bank, and a central controller. The cache management module includes a local controller and a cache memory in which operands are stored for calculating future twiddle factors. The modular multipliers bank includes an interconnection matrix at the input distributing operands on the modular multiplier inputs. The circuit can be configured to minimise the size of the cache memory and/or reduce the latency of the twiddle factor sequence calculation. Finally, the generating circuit may include several calculation management modules sharing the same modular multipliers bank to generate sequences of twiddle factors on several finite fields.

Abstract: The present invention relates to a stream-based NTT processor comprising: a plurality (K) of processing stages (210k, k=0, . . . , K?1) organised in a pipeline (210); a plurality (G+1) of memory banks (220g, g=0, . . . , G); a read management module (260) for reading, within one memory) of a memory bank (220g) of the processor, sets of twiddle factors intended for parameterising a processing stage (210k); a write management module (270) for receiving, in the form of successive blocks, a set of twiddle factors and writing said sets of twiddle factors into the memories of a memory bank, the writing being carried out cyclically in the memory banks, each new set of twiddle factors being written into a new memory bank; and a control module for controlling the writing and reading of twiddle factors as well as the progression of data blocks through the processing stages.

Abstract: A method for federative learning of an artificial neural network model on a plurality of sets of training data. The learning method used a plurality of data suppliers each having a distinct set of training data and a symmetric key, and an aggregation platform aggregating partial models trained on a sub-plurality of these sets, in each iteration. In each iteration, the platform selects a sub-plurality of data suppliers and supplies the parameters of the model to them, in homomorphically encrypted form. Each training data supplier decrypts these parameters, trains the model on his own data set and returns the parameters of the partial model thus obtained, encrypted by his symmetric key, to the platform. The aggregation platform then transcrypts these partial models in the homomorphic domain, and combines them in the homomorphic domain to obtain a global model.

Abstract: A method for confidentially querying an encrypted database hosted by a server. The user transmits, to the server, a request including a predicate. The predicate is evaluated in an encrypted manner on different records of the database. The records that satisfy the query are transferred blindly into a container. The container is transmitted to the user who decrypts the content of same. If the container is full, the user sends a continuation request to the server. If this is not the case, the records of the successive containers that have already been decrypted form the response to the request.

Abstract: A method allowing a user confidentially to query a server including a database. The user builds, from an index of a desired entry, a vector encrypted by a homomorphic probabilistic cryptosystem, and transmits a request having the vector as an argument to the server. The server performs a scalar product between the vector of entries of the database and returns the result to the user. The user decrypts the scalar product to obtain the desired entry. The method can be used for example for querying a location-based service and for navigation by concealed routes.

Abstract: The present invention concerns a method for secure data classification by a computer platform. A client sends to the platform data to be classified in encrypted form using a first symmetric key. Similarly, a supplier sends to the platform parameters of a classification model in encrypted form using a second symmetric key. The invention uses a homomorphic cryptosystem defined by a public key and a private key. The platform performs a first transcryption step by deciphering the data to be classified in the homomorphic domain and a second transcryption step by deciphering the model parameters in the homomorphic domain. The classification function is then evaluated in the homomorphic domain for providing a classification result encrypted by said public key.

Abstract: A device and a method for encrypting a virtual machine by on-the-fly encryption and decryption of the memory is described. The device comprises hardware and software elements, including a hardware security module for executing, in response to data read/write requests, on-the-fly encryption and decryption operations of the data stream associated with an encrypted memory image. The device and the method are particularly suited to allowing the secure authentication and execution of programs in the context of the infrastructure as a service layer of the cloud computing model.

Abstract: A method for confidentially querying an encrypted database hosted by a server. The user transmits, to the server, a request including a predicate. The predicate is evaluated in an encrypted manner on different records of the database. The records that satisfy the query are transferred blindly into a container. The container is transmitted to the user who decrypts the content of same. If the container is full, the user sends a continuation request to the server. If this is not the case, the records of the successive containers that have already been decrypted form the response to the request.

Abstract: A system for executing code with blind hypervision mechanism comprises: at least one addressable physical memory, a processor operating in at least two modes, a mode termed initialization making it possible to define at least one partition in the memory and at least one second mode termed nominal, a memory bus linking the processor to the memory, a memory partitioning unit positioned on the memory bus, the unit being adapted for restricting memory access to the partition currently executing when the processor is in a mode other than the initialization mode.

Abstract: A method allowing a user confidentially to query a server including a database. The user builds, from an index of a desired entry, a vector encrypted by a homomorphic probabilistic cryptosystem, and transmits a request having the vector as an argument to the server. The server performs a scalar product between the vector of entries of the database and returns the result to the user. The user decrypts the scalar product to obtain the desired entry. The method can be used for example for querying a location-based service and for navigation by concealed routes.

Abstract: A system for executing code with blind hypervision mechanism comprises: at least one addressable physical memory, a processor operating in at least two modes, a mode termed initialization making it possible to define at least one partition in the memory and at least one second mode termed nominal, a memory bus linking the processor to the memory, a memory partitioning unit positioned on the memory bus, the unit being adapted for restricting memory access to the partition currently executing when the processor is in a mode other than the initialization mode.

Abstract: A device and a method for encrypting a virtual machine by on-the-fly encryption and decryption of the memory is described. The device comprises hardware and software elements, including a hardware security module for executing, in response to data read/write requests, on-the-fly encryption and decryption operations of the data stream associated with an encrypted memory image. The device and the method are particularly suited to allowing the secure authentication and execution of programs in the context of the infrastructure as a service layer of the cloud computing model.