Documents & Deliverables
Secure Risk Propagation
Large sums of cash or cryptocurrencies are often scattered throughout multiple accounts and transferred via several hops to a central account. The secure risk propagation algorithm allows to propagate these types of risks throughout a multi-bank transaction network.
The TNO PET Lab is a cross-project initiative initiated to improve the overall quality, generality, and reusability in the development of secure Multi-Party Computation (MPC) solutions developed in the numerous (past, ongoing, and future) TNO projects that involve MPC.
Privacy preserving graph embeddings
Compared to other data structures, transaction graphs are rich in information: they contain information about bank accounts, transactions, local and global structure information. Common machine learning methods cannot easily handle such structured data and that is where graph embeddings come into play. They circumvent this issue by generating a graph embedding for each node of the graph.
Privacy-preserving analytics for secure collaborative detection of financial crime
In the fight against financial crime, collaboration between banks is key. At the same time, protection of customer privacy is more important than ever. Banks are looking for ways to balance between privacy and value in their detection techniques. Privacy-enhancing cryptographic techniques, which lately have been moving from academics to real-world applications, can help maintaining this balance. These techniques enable collaborative analysis on sensitive data, while keeping the confidential and personal data private.