Hi! I am a Ph.D. student at Seoul National University. My advisor is Prof. Jung Hee Cheon. I am interested in a broad range of topics in cryptography. My current research focuses on homomorphic encryption and its applications.
Imagine training a machine learning model on sensitive medical data—like patient records—without ever seeing the actual data. How can we ensure privacy while still allowing useful computation?
This is the core idea behind Privacy-Preserving Machine Learning (PPML), and Fully Homomorphic Encryption (FHE) plays a central role. FHE allows computations to be performed directly on encrypted data, without needing to decrypt it first. The results, when decrypted, match what would have been obtained using the original plaintext.
Mathematically, homomorphic encryption schemes are designed so that the decryption function is a homomorphism with respect to certain operations—like addition and multiplication. This means operations on ciphertexts translate directly into corresponding operations on plaintexts.
This capability enables secure data processing in areas like healthcare, finance, and artificial intelligence, where data confidentiality is critical. With Fully Homomorphic Encryption (FHE), we can achieve both functionality and privacy, without compromise.
