Secure calculation: calculating on encrypted or private data

Cryptography (encryption, signatures, etc.) is used to guarantee the confidentiality and integrity of data at rest (stored in files and databases) or in transit (during communications over networks). However, to compute on this data, it has traditionally been necessary to decrypt it and reveal it entirely to the computation operator, with the attendant risks of data "leakage". Recent advances in cryptography and protocols have led to the emergence of alternative approaches in which data remains encrypted or private during computation, thus guaranteeing its confidentiality. This lecture and the accompanying seminar provide an introduction to the issue of secure computing. They present recent results in this field: secure multiparty computation, homomorphic encryption, functional encryption, cryptographic obfuscation, etc., as well as some of their applications, notably for electronic voting.