Latest recommendation systems show immense potential in data protection. Privacy is preserved when searching for content, while providers can offer their service with consistent quality and performance.
Online services such as Amazon, Netflix or Spotify use recommendation systems as the basis of their success. Recommendation systems collect data about users and make recommendations based on this data. This makes it easier to navigate through the flood of information, but it also raises questions about data security and the protection of privacy.
Recommendation systems developed by the Know Center are novel in that they preserve privacy and data protection in their operations. For this purpose, they work with completely encrypted data from start to finish.
The recommender systems have no direct insight into the data and “calculate” with encrypted values, but they can still generate targeted recommendations tailored to the individual user.
In addition to encryption methods, minimization techniques are used to guarantee maximum performance. The systems can query 100 million data entries within 10 seconds and provide high-quality recommendations. This opens the door to many real-world applications in the field of privacy-friendly recommendation systems.