Which technique enables performing computations on encrypted data without decrypting?

Which technique enables performing computations on encrypted data without decrypting?

• A. Data augmentation
• B. Homomorphic encryption
• C. Secure enclaves
• D. Differential privacy

Answer: (B)

Performing computations on encrypted data without decrypting is achieved with homomorphic encryption. This cryptographic approach lets you apply operations—such as addition or multiplication—directly to ciphertexts. When you later decrypt the result, you get the same outcome as if you had performed the operations on the plaintext, enabling privacy-preserving processing in untrusted environments like the cloud. There are variations: some schemes support specific operations (partially homomorphic), while fully homomorphic encryption supports arbitrary computations on encrypted data. Data augmentation involves creating more data samples for training and has nothing to do with encryption. Secure enclaves provide a protected execution environment where data can be processed securely, but the data is decrypted inside the hardware during computation. Differential privacy adds noise to data or results to protect individual identities, rather than enabling computation on encrypted data.

Performing computations on encrypted data without decrypting is achieved with homomorphic encryption. This cryptographic approach lets you apply operations—such as addition or multiplication—directly to ciphertexts. When you later decrypt the result, you get the same outcome as if you had performed the operations on the plaintext, enabling privacy-preserving processing in untrusted environments like the cloud. There are variations: some schemes support specific operations (partially homomorphic), while fully homomorphic encryption supports arbitrary computations on encrypted data.

Data augmentation involves creating more data samples for training and has nothing to do with encryption. Secure enclaves provide a protected execution environment where data can be processed securely, but the data is decrypted inside the hardware during computation. Differential privacy adds noise to data or results to protect individual identities, rather than enabling computation on encrypted data.