Recently, a research team led by Dr. Yuncong Hu, an Assistant Professor at SJTU and a Zhiyuan Computer Science Alumni (Class of 2017) has proposed "HasteBoots," a protocol that slashes the proof generation time for Fully Homomorphic Encryption (FHE) bootstrapping from hours to seconds. The paper, authored by Carnegie Mellon Ph.D. student Liu Fengrun and supervised by Hu, has been accepted to USENIX Security '26.

FHE enables computation on encrypted data but faces a verification bottleneck, particularly during bootstrapping. HasteBoots addresses this by designing efficient quotient-ring arithmetic protocols for TFHE, achieving proof generation in seconds for a single evaluation and under one minute for batching 16 operations. Verification remains ultra-fast at 126 milliseconds with a proof size of just 0.28 MB, significantly outperforming prior art. This breakthrough removes a key barrier for verifiable outsourced computation and privacy-preserving machine learning.

Dr. Hu, a UC Berkeley Ph.D. graduate and a national young talent awardee, is renowned for foundational work like the Marlin algorithm and Arkworks toolkit. HasteBoots represents another step in his mission to bridge cutting-edge cryptography with practical applications. Another of his recent works, DFS, has also been accepted to USENIX Security '25.