ZEro-Trust Attack (ZETA) framework with Split Learning for Autonomous Vehicles in 6G Networks
.jpg)
In this paper, we propose a ZEro-Trust Attack (ZETA) framework for data reconstruction and model inversion attacks for autonomous vehicles opting for split learning strategies. The model inversion attacks aim to attack trained models for reconstructing of data that could violate the privacy of the users. We propose the joint training of client, server, and shadow models for both the reconstruction and main task to fool existing methods such as Gradient Scrutinizer. This study is proposed as a basis to design more sophisticated defence mechanism for autonomous vehicles to protect user services in 5G/6G networks.
The paper was accepted in IEEE Workshop on 6G for Connected and Immersive Intelligence at IEEE WCNC (a flagship conference) and endorsed by one6G Association.
IEEE Wireless Communications and Networking Conference (WCNC) is the world premiere wireless event that brings together industry professionals, academics, and individuals from government agencies and other institutions to exchange information and ideas on wireless advancement communications and networking technology. Among all the conferences in the world (IEEE and otherwise) in the broad area of electronics and electrical engineering, WCNC ranks #7 among 773 conferences according to the impact score on Research.com.
The workshop was supported by one6G (https://one6g.org/) that aims to become a forum to address technological research in support of the usage scenarios defined for the IMT-2030 vision, namely, (i) immersive communication, massive communication, and hyper-reliable and low-latency communication). (ii) coverage extension (e.g., ubiquitous connectivity), and (iii) service extension beyond communication.
