To derive valuable insights from statistics, machine learning applications frequently analyze substantial amounts of data. In this work, we address the problem of designing efficient secure techniques to probe large datasets which allow a scientist to conduct large-scale medical studies over specific attributes of patients' records, while maintaining the privacy of his model. We introduce a set of composable homomorphic operations and show how to combine private functions evaluation with private thresholds via approximate fully homomorphic encryption. This allows us to design a new system named TETRIS, which solves the real-world use case of private functional exploration of large databases, where the statistical criteria remain private to the server owning the patients’ records. Our experiments show that TETRIS achieves practical performance over a large dataset of patients even for the evaluation of elaborate statements composed of linear and nonlinear functions. It is possible to extract private insights from a database of hundreds of thousands of patient records within only a few minutes on a single thread, with an amortized time per database entry smaller than 2ms.

Principal component analysis (PCA) is an essential algorithm for dimensionality reduction in many data science domains. We address the problem of performing a federated PCA on private data distributed among multiple data providers while ensuring data confidentiality. Our solution, SF-PCA, is an end-to-end secure system that preserves the confidentiality of both the original data and all intermediate results in a passive-adversary model with up to all-but-one colluding parties. SF-PCA jointly leverages multiparty homomorphic encryption, interactive protocols, and edge computing to efficiently interleave computations on local cleartext data with operations on collectively encrypted data. SF-PCA obtains results as accurate as non-secure centralized solutions, independently of the data distribution among the parties. It scales linearly or better with the dataset dimensions and with the number of data providers. SF-PCA is more precise than existing approaches that approximate the solution by combining local analysis results, and between 3x and 250x faster than privacy-preserving alternatives based solely on secure multiparty computation or homomorphic encryption. Our work demonstrates the practical applicability of secure and federated PCA on private distributed datasets.

We present RHODE, a novel system that enables privacy-preserving training of and prediction on Recurrent Neural Networks (RNNs) in a cross-silo federated learning setting by relying on multiparty homomorphic encryption. RHODE preserves the confidentiality of the training data, the model, and the prediction data; and it mitigates federated learning attacks that target the gradients under a passive-adversary threat model. We propose a packing scheme, multi-dimensional packing, for a better utilization of Single Instruction, Multiple Data (SIMD) operations under encryption. With multi-dimensional packing, RHODE enables the efficient processing, in parallel, of a batch of samples. To avoid the exploding gradients problem, RHODE provides several clipping approximations for performing gradient clipping under encryption. We experimentally show that the model performance with RHODE remains similar to non-secure solutions both for homogeneous and heterogeneous data distributions among the data holders. Our experimental evaluation shows that RHODE scales linearly with the number of data holders and the number of timesteps, sub-linearly and sub-quadratically with the number of features and the number of hidden units of RNNs, respectively. To the best of our knowledge, RHODE is the first system that provides the building blocks for the training of RNNs and its variants, under encryption in a federated learning setting.

Abstract In this paper, we address the problem of privacy-preserving distributed learning and the evaluation of machine-learning models by analyzing it in the widespread MapReduce abstraction that we extend with privacy constraints. We designspindle(Scalable Privacy-preservINg Distributed LEarning), the first distributed and privacy-preserving system that covers the complete ML workflow by enabling the execution of a cooperative gradient-descent and the evaluation of the obtained model and by preserving data and model confidentiality in a passive-adversary model with up to N −1 colluding parties.spindleuses multiparty homomorphic encryption to execute parallel high-depth computations on encrypted data without significant overhead. We instantiatespindlefor the training and evaluation of generalized linear models on distributed datasets and show that it is able to accurately (on par with non-secure centrally-trained models) and efficiently (due to a multi-level parallelization of the computations) train models that require a high number of iterations on large input data with thousands of features, distributed among hundreds of data providers. For instance, it trains a logistic-regression model on a dataset of one million samples with 32 features distributed among 160 data providers in less than three minutes.

Abstract We propose and evaluate a secure-multiparty-computation (MPC) solution in the semi-honest model with dishonest majority that is based on multiparty homomorphic encryption (MHE). To support our solution, we introduce a multiparty version of the Brakerski-Fan-Vercauteren homomorphic cryptosystem and implement it in an open-source library. MHE-based MPC solutions have several advantages: Their transcript is public, their o~ine phase is compact, and their circuit-evaluation procedure is noninteractive. By exploiting these properties, the communication complexity of MPC tasks is reduced from quadratic to linear in the number of parties, thus enabling secure computation among potentially thousands of parties and in a broad variety of computing paradigms, from the traditional peer-to-peer setting to cloud-outsourcing and smart-contract technologies. MHE-based approaches can also outperform the state-of-the-art solutions, even for a small number of parties. We demonstrate this for three circuits: private input selection with application to private-information retrieval, component-wise vector multiplication with application to private-set intersection, and Beaver multiplication triples generation. For the first circuit, privately selecting one input among eight thousand parties’ (of 32 KB each) requires only 1.31 MB of communication per party and completes in 61.7 seconds. For the second circuit with eight parties, our approach is 8.6 times faster and requires 39.3 times less communication than the current methods. For the third circuit and ten parties, our approach generates 20 times more triples per second while requiring 136 times less communication per-triple than an approach based on oblivious transfer. We implemented our scheme in the Lattigo library and open-sourced the code at github.com/ldsec/lattigo.