This page provides access to all public deliverables submitted by SODA.
This deliverable contains research results towards the application of MPC to large volumes of data: new results on secure linear algebra, protocols in the streaming setting, new MPC protocols for (obliviously) applying a function to certain elements of a sparse data set, with applications to Kaplan–Meier survival analysis, on combining two-party computation with private set intersection, and obliviously training a decision tree on a secret-shared database with continuous attributes.
This deliverable presents research on designing and improving special-purpose protocols for dedicated secure computation tasks related to data analytics. Firstly, we present optimizations of multi-party computation (MPC) protocols over rings. We also present efficient protocols for neural network evaluation and ridge regression, scaling to large datasets with millions of records. Finally, we look at a new approach to carrying out the preprocessing phase of MPC protocols, which drastically reduces communication.
This deliverable presents research on differential privacy and leakage control in the context of secure multi-party computation (MPC). The main goal of this research is to obtain a better understanding of information leakage that can occur as a result of taking part in a secure computation, and investigate practical techniques that can be used to mitigate this.
This deliverable presents research on general-purpose protocols for secure computation. The main goal of this research is to develop new, optimized protocols and primitives that can be used for applications requiring private data analytics, particularly when the quantity of data becomes large.
This deliverable presents research results on the topic of application oriented secure multiparty computation with a focus on data mining applications.
This report summarizes the deliverable D4.2 proof-of-concept implementations developed in the first phase of the SODA project. It touches on MPC advancements in the FRESCO framework, developer-friendly aspects of MPC frameworks such as FRESCO and MPyC, verifiable computation using zk-SNARKs and implementation of data analytics and machine learning algorithms.
This deliverable contains state of the art analysis in secure multi-party computation (MPC), which includes a survey of both theoretical results and practical implementations. The goal of this deliverable is to provide an overview of the best available techniques in the state-of-the-art as opposed to a complete historic account of all known techniques. It is part of Work Package 1 of the SODA project.
This deliverable contains a state of the art analysis in MPC-based privacy-preserving data mining protocols for use with Big Data.
This deliverable intends to provide a thorough legal analysis of the current privacy law in the EU, with emphasis placed on the GDPR, which shall apply from the 25th May 2018. Special emphasis is on regulations related to the utilisation of Big Data, as well as on the genuine conflict between Big Data and privacy.
This contains a plan for the user studies to be conducted in the SODA project.
Summary: A public website for the SODA project, https://soda-project.eu, has been set up, currently containing a project summary and description; overview of the consortium; information on and links to the websites of all project members; project news; lists of publications and deliverables with download links, when available; information about impact; and contact information.