i-SEAL²: Identifying spam EmAiL with SEAL

End-to-end encrypted emails are desirable with regards to privacy, as it prevents your email provider from storing and reading your emails in plaintext. However, with the perk of privacy from the end-to-end encryption, you lose the spam filter, as the filtering process requires an analysis on the email's content, or its metadata. The classification of whether an email is spam typically relies on machine learning algorithms that have been trained on large amounts of emails. A naive approach to combine end-to-end encryption of emails and a spam filter would be for every user to simply build their own model using only their own emails to train the machine learning model. However, one user typically only has a limited number of emails and this local approach is going to result in a model which is less accurate than the one provided by an email provider, simply due to the size of the dataset used to train the machine learning model. In order to obtain an accurate model, large amounts of diverse data are required.