nehatirumalasetti / fhe-workspace

Privacy protected recommendation engine based on social media user profiling using IBM FHE Toolkit.

In traditional encryption schemes, data remains in an encrypted format in transit and storage but needs to be decrypted to perform any computation on it. Homomorphic Encryption is a technique that allows us to process data without having to decrypt it. Thus, the results are decrypted at the user’s end directly, and are kept concealed from the system. Currently, sensitive user data is mined to collect insights for recommendations such as friend suggestions, ads, etc.Our system provides a proof of concept that demonstrates the application of fully homomorphic encryption(FHE) in the social networking domain using the IBM FHE toolkit. In our project, we demonstrate the concept of homomorphic encryption by processing the data of Twitter users, building user profiles and recommending similar users to follow for Twitter users, all while the data remains encrypted. This ensures that in case of any data breach, the user’s privacy is not violated, and user data and their insights are protected from the party generating the recommendations as well.