Operations on Encrypted Data is one of the core components of the ASCLEPIOS vision. With this as a starting point of our research, we try to push the boundaries of traditional machine learning algorithms, by designing models that are applied directly on encrypted data.
Tanveer Khan (TUNI) presented the paper titled “Blind Faith: Privacy-Preserving Machine Learning Using Function Approximation” that won the Student’s Best Paper Award, at ISCC 2021!
This work proposes a machine learning model – “Blind Faith”– in which the training phase occurs in plaintext data, but the classification of the users’ inputs is performed on homomorphically encrypted ciphertexts (Figure 1), an approach offering a solution for the deployment of machine learning models on remote untrusted clouds while preserving the users’ privacy.
The IEEE Symposium on Computers and Communications (ISCC) provides an international technical forum for experts from industry and academia to exchange ideas and present results of on-going research in most state-of-the-art areas of computers and communications. In its 26th edition, ISCC 2021 provided an insight into the unique world stemming from the interaction between the fields of computers and communications. ISCC 2021 took place in Athens, Greece (and also as a virtual event), on September 5-8, 2021.