Ecole Polytechnique Federale De Lausanne

Ecole Polytechnique Féderale De Lausanne (Epfl)

Sinem Sav

Jean-Pierre Hubaux

David Froelicher

Jean-Philippe Bossuat

Apostolos Pyrgelis

Juan Ramon Troncoso-Pastoriza

Joao Gomes De Sa E Sousa

Joao Gomes De SÁ E Sousa

A computer-implemented method and a distributed computer system for privacy-preserving distributed training of a global neural network model on distributed datasets. The system has data providers communicatively coupled with each having a respective local training dataset and a vector of output labels for training the global model. Further, it has a cryptographic distributed secret key and a corresponding collective cryptographic public key of a multiparty fully homomorphic encryption scheme, with the weights of the global model being encrypted with the collective public key. Each data provider computes and aggregates, for each layer of the global model, encrypted local gradients using the respective local training dataset and output labels, with forward pass and backpropagation using stochastic gradient descent. One data provider homomorphically combines the current local gradients of the data providers into combined local gradients, and updates the weights of the current global model based on the combined local gradients.

System and method for privacy-preserving distributed training of neural network models on distributed datasets

A system for privacy-preserving distributed training of a global model on distributed datasets has a plurality of data providers being communicatively coupled. Each data provider has a local model and a local training dataset for training the local model using an iterative training algorithm. Further it has a portion of a cryptographic distributed secret key and a corresponding collective cryptographic public key of a multiparty fully homomorphic encryption scheme. All models are encrypted with the collective public key. Each data provider trains its local model using the respective local training dataset, and combines the local model with the current global model into a current local model. A data provider homomorphically combines current local models into a combined model, and updates the current global model based on the combined model. The updated global model is provided to at least a subset of the other data providers.

System and method for privacy-preserving distributed training of machine learning models on distributed datasets

Growing community of inventors

Lausanne, Switzerland

Sinem Sav