TY - GEN
T1 - MediSC: Towards Secure and Lightweight Deep Learning as a Medical Diagnostic Service
AU - Liu, Xiaoning
AU - Zheng, Yifeng
AU - Yuan, Xingliang
AU - Yi, Xun
N1 - Publisher Copyright:
© 2021, Springer Nature Switzerland AG.
PY - 2021/9
Y1 - 2021/9
N2 - The striking progress of deep learning paves the way towards intelligent and quality medical diagnostic services. Enterprises deploy such services via the neural network (NN) inference, yet confronted with rising privacy concerns of the medical data being diagnosed and the pre-trained NN models. We propose, a system framework that enables enterprises to offer secure medical diagnostic service to their customers via an execution of NN inference in the ciphertext domain. ensures the privacy of both parties with cryptographic guarantees. At the heart, we present an efficient and communication-optimized secure inference protocol that purely relies on the lightweight secret sharing techniques and can well cope with the commonly-used linear and non-linear NN layers. Compared to the garbled circuits based solutions, the latency and communication of are 24 × lower and 868 × less for the secure ReLU, and 20 × lower and 314 × less for the secure Max-pool. We evaluate on two benchmark and four real-world medical datasets, and comprehensively compare it with prior arts. The results demonstrate the promising performance of, which is much more bandwidth-efficient compared to prior works.
AB - The striking progress of deep learning paves the way towards intelligent and quality medical diagnostic services. Enterprises deploy such services via the neural network (NN) inference, yet confronted with rising privacy concerns of the medical data being diagnosed and the pre-trained NN models. We propose, a system framework that enables enterprises to offer secure medical diagnostic service to their customers via an execution of NN inference in the ciphertext domain. ensures the privacy of both parties with cryptographic guarantees. At the heart, we present an efficient and communication-optimized secure inference protocol that purely relies on the lightweight secret sharing techniques and can well cope with the commonly-used linear and non-linear NN layers. Compared to the garbled circuits based solutions, the latency and communication of are 24 × lower and 868 × less for the secure ReLU, and 20 × lower and 314 × less for the secure Max-pool. We evaluate on two benchmark and four real-world medical datasets, and comprehensively compare it with prior arts. The results demonstrate the promising performance of, which is much more bandwidth-efficient compared to prior works.
KW - Neural network inference
KW - Privacy-preserving medical service
KW - Secret sharing
KW - Secure computation
UR - http://www.scopus.com/inward/record.url?scp=85116876085&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-88418-5_25
DO - 10.1007/978-3-030-88418-5_25
M3 - Conference article published in proceeding or book
AN - SCOPUS:85116876085
SN - 9783030884178
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 519
EP - 541
BT - Computer Security – ESORICS 2021 - 26th European Symposium on Research in Computer Security, Proceedings
A2 - Bertino, Elisa
A2 - Shulman, Haya
A2 - Waidner, Michael
PB - Springer Science and Business Media Deutschland GmbH
T2 - 26th European Symposium on Research in Computer Security, ESORICS 2021
Y2 - 4 October 2021 through 8 October 2021
ER -