Periodic Reporting for period 1 - CryptAnalytics (Secure Analytics as Business Services)
Période du rapport: 2021-04-01 au 2022-09-30
Big data and data lakes are gold mines for data scientists with tons of applications to finance, medicine, economics, etc. But most of these data are quite sensitive and cannot be widely distributed or even just used without strong protection. One is thus facing the huge dilemma of having to make the choice between highly valuable analytics and privacy. The ERC CryptoCloud project has designed several primitives perfectly well-suited to such situations. They are quite powerful but also highly technical for appropriate deployment at scale. The European start-up Cosmian has been interested in our tools, and we have already started a fruitful collaboration.
The main goal of the CryptAnalytics project is to develop a unified platform that will allow dealing with many situations, for secure analytics as business services. This will accelerate the development of Cosmian, giving a major competitive advantage on the fast emerging market of data analytics, with strong privacy guarantees by design. Companies are realizing the incredible value of privacy-preserving analytics. As a result, the number of business use-cases is exploding, but each application is a new challenge in itself, with new kinds of evaluations, new privacy concerns, and different amounts of data.
More precisely, a prototype has been developed to provide benchmarks for the main tool DMCFE (Decentralized Multi-Client Functional Encryption) proposed by the ERC CryptoCloud, and two new constructions have been proposed, to appear at ACM CCS '22 and IACR Asiacrypt '22.
The main goal of the CryptAnalytics project is to develop a unified platform that will allow dealing with many situations, for secure analytics as business services. This will accelerate the development of Cosmian, giving a major competitive advantage on the fast emerging market of data analytics, with strong privacy guarantees by design. Companies are realizing the incredible value of privacy-preserving analytics. As a result, the number of business use-cases is exploding, but each application is a new challenge in itself, with new kinds of evaluations, new privacy concerns, and different amounts of data.
More precisely, a prototype has been developed to provide benchmarks for the main tool DMCFE (Decentralized Multi-Client Functional Encryption) proposed by the ERC CryptoCloud, and two new constructions have been proposed, to appear at ACM CCS '22 and IACR Asiacrypt '22.