We unlock the power of sensitive data collaborations for you to make better decisions.
Your data remains, all of the time, in your control. No one else has access to it. With advanced collective analytics and federated learning extracted from all participating organizations, we provide you with a new kind of insights so that you can make better decisions.
Tune Insight is a startup incubated at the EPFL Laboratory for Data Security. We are currently developing proofs-of-concept and deploying prototypes with early customers in healthcare, e.g. with Swiss university hospitals (MedCo), in cyber threat intelligence, in insurance risk analysis and in other domains where data-driven decision making can be improved by secure collaborations on sensitive data.
As organizations become more data driven, they realize that their own data is not sufficient to make sound decisions, but they are prevented from entering sensitive data collaborations with others due to fear of data leaks, or because of data-privacy concerns and regulations.
Tune Insight's system is distributed, with software deployed at each connected organization, close to their respective sensitive data
An authorized user from a participating organization defines and launches a query
Tune Insight software at each connected organization then prepares encrypted partial results without sharing or transferring the organization's data (not even to Tune Insight)
Tune Insight's distributed system aggregates all results and sends back encrypted powerful insight that only the authorized user can decrypt and use to make better decisions
Bringing you the latest in privacy and encryption technologies.
Built on state-of-the-art privacy and encryption technologies emerging from years of world-class research, our software solutions feature a highly efficient combination of secure multiparty computation and quantum-resistant homomorphic encryption, that we term multi-party homomorphic encryption (see Multiparty Homomorphic Encryption from Ring-Learning-With-Errors and Lattigo for details).
GDPR compliant, more secure than federated learning, not moving or revealing your data.
More scalable than traditional multi-party computation. Manage hundreds of participants.
As more organizations join the collaboration, insights get more precise and valuable.
Highly parallelizable, we bring better performance than fully homomorphic encryption.
Fully software-based, no need for trusted 3rd party hardware or new infrastructure.