eduzhai > Applied Sciences > Engineering >

Adversarial Machine Learning based Partial-model Attack in IoT

  • Save

... pages left unread,continue reading

Document pages: 6 pages

Abstract: As Internet of Things (IoT) has emerged as the next logical stage of theInternet, it has become imperative to understand the vulnerabilities of the IoTsystems when supporting diverse applications. Because machine learning has beenapplied in many IoT systems, the security implications of machine learning needto be studied following an adversarial machine learning approach. In thispaper, we propose an adversarial machine learning based partial-model attack inthe data fusion aggregation process of IoT by only controlling a small part ofthe sensing devices. Our numerical results demonstrate the feasibility of thisattack to disrupt the decision making in data fusion with limited control ofIoT devices, e.g., the attack success rate reaches 83 when the adversarytampers with only 8 out of 20 IoT devices. These results show that the machinelearning engine of IoT system is highly vulnerable to attacks even when theadversary manipulates a small portion of IoT devices, and the outcome of theseattacks severely disrupts IoT system operations.

Please select stars to rate!


0 comments Sign in to leave a comment.

    Data loading, please wait...