Privacy-preserving computation is a family of techniques that allow data to be processed or analysed without exposing the underlying plaintext to the computing party. It includes homomorphic encryption, secure multiparty computation, trusted execution environments, federated learning, and differential privacy. These methods enable collaboration and analytics over sensitive data while maintaining confidentiality.
Content
- Homomorphic encryption permits arithmetic on ciphertexts; secure multiparty computation splits computation across parties so none sees the whole input; trusted execution environments isolate computation in hardware enclaves; and federated learning trains models on-device, sharing only updates. Differential privacy adds calibrated noise to bound what any output reveals about an individual. These approaches trade additional computation or accuracy for strong confidentiality guarantees.