Commissariat a L'energie Atomique Et Aux Energies Alternatives

Sergiu Carpov

Renaud Sirdey

Oana Stan

Simon Fau

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.

Low latency calculation transcryption method

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 iof the class to which it belongs, is returned to the user.

Method and system for confidential classification of data

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.

Method for secure classification using a transcryption operation

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.

Method for collaborative learning of an artificial neural network without disclosing training data

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.

Growing community of inventors

Massy, France

Sergiu Carpov