One Paper Accepted by CCS'2026
One paper on privacy-preserving AI is accepted by CCS'2026.
The title of the paper is “ROSETTA: Efficient and Accurate Privacy-Preserving LLM Decoding via Hybrid CKKS/TFHE Evaluation”, featuring a new way to evaluate the nonlinear functions in LLMs through hybrid CKKS/TFHE schemes. This is led by our PhD student Jiangrui Yu.