PrivaScissors: Enhance the Privacy of Collaborative Inference through the Lens of Mutual Information
원문 초록
Edge-cloud collaborative inference empowers resource-limited IoT devices to support deep learning applications without disclosing their raw data to the cloud server, thus preserving privacy. Nevertheless, prior research has shown that collaborative inference still results in the exposure of data and predictions from edge devices. To enhance the privacy of collaborative inference, we introduce a defense strategy called PrivaScissors, which is designed to reduce the mutual information between a model's intermediate outcomes and the device's data and predictions. We evaluate PrivaScissors's performance on several datasets in the context of diverse attacks and offer a theoretical robustness guarantee.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Theoretical Insights in Model Inversion Robustness and Conditional Entropy Maximization for Collaborative Inference Systems
PhyloVAE: Unsupervised Learning of Phylogenetic Trees via Variational Autoencoders
CITER: Collaborative Inference for Efficient Large Language Model Decoding with Token-Level Routing
Distributed Mixture-of-Agents for Edge Inference with Large Language Models
GREAT: Geometry-Intention Collaborative Inference for Open-Vocabulary 3D Object Affordance Grounding
Private Collaborative Edge Inference via Over-the-Air Computation
토론
이 논문과 연결된 포럼 글이 아직 없습니다.