Joint Optimization of Prompt Security and System Performance in Edge-Cloud LLM Systems

doi:10.1109/INFOCOM55648.2025.11044720

科研成果详情

题名	Joint Optimization of Prompt Security and System Performance in Edge-Cloud LLM Systems
作者	Huang，Haiyang 1; Meng，Tianhui 2; Jia，Weijia 1,2
发表日期	2025
会议名称	2025 IEEE Conference on Computer Communications, INFOCOM 2025
会议录名称	Proceedings - IEEE INFOCOM
ISSN	0743-166X
会议日期	2025-05-19——2025-05-22
会议地点	London
摘要	Large language models (LLMs) have significantly facilitated human life, and prompt engineering has improved the efficiency of these models. However, recent years have witnessed a rise in prompt engineering-empowered attacks, leading to issues such as privacy leaks, increased latency, and system resource wastage. Though safety fine-tuning based methods with Reinforcement Learning from Human Feedback (RLHF) are proposed to align the LLMs, existing security mechanisms fail to cope with fickle prompt attacks, highlighting the necessity of performing security detection on prompts. In this paper, we jointly consider prompt security, service latency, and system resource optimization in Edge-Cloud LLM (EC-LLM) systems under various prompt attacks. To enhance prompt security, a vector-database-enabled lightweight attack detector is proposed. We formalize the problem of joint prompt detection, latency, and resource optimization into a multi-stage dynamic Bayesian game model. The equilibrium strategy is determined by predicting the number of malicious tasks and updating beliefs at each stage through Bayesian updates. The proposed scheme is evaluated on a real implemented EC-LLM system, and the results demonstrate that our approach offers enhanced security, reduces the service latency for benign users, and decreases system resource consumption compared to state-of-the-art algorithms.
关键词	Bayesian game edge-cloud LLM Prompt attack resource optimization
DOI	10.1109/INFOCOM55648.2025.11044720
URL	查看来源
语种	英语English
Scopus入藏号	2-s2.0-105011085385
引用统计
文献类型	会议论文
条目标识符	https://repository.uic.edu.cn/handle/39GCC9TT/13735
专题	北师香港浸会大学
通讯作者	Meng，Tianhui; Jia，Weijia
作者单位	1.Institute of Artificial Intelligence and Future Networks,Beijing Normal University,Zhuhai,China 2.BNU-HKBU United International College,Department of Computer Science,Zhuhai,China
通讯作者单位	北师香港浸会大学
推荐引用方式 GB/T 7714	Huang，Haiyang,Meng，Tianhui,Jia，Weijia. Joint Optimization of Prompt Security and System Performance in Edge-Cloud LLM Systems[C], 2025.