PROJECT TITLE:

Insider Collusion Attack on Privacy-Preserving Kernel-Based Data Mining Systems - 2016

ABSTRACT:

In this paper, we contemplate a new insider threat for the privacy preserving work of distributed kernel-based mostly information mining (DKBDM), such as distributed support vector machine. Among many known information breaching problems, those related to insider attacks have been rising significantly, creating this one of the fastest growing sorts of security breaches. Once considered a negligible concern, insider attacks have risen to be one amongst the top 3 central information violations. Insider-related analysis involving the distribution of kernel-based knowledge mining is limited, resulting in substantial vulnerabilities in coming up with protection against collaborative organizations. Prior works typically fall short by addressing a multifactorial model that's more restricted in scope and implementation than addressing insiders within a company colluding with outsiders. A faulty system allows collusion to travel unnoticed when an insider shares data with an outsider, who will then recover the original information from message transmissions (intermediary kernel values) among organizations. This attack requires solely accessibility to a few knowledge entries inside the organizations instead of requiring the encrypted administrative privileges typically found within the distribution of data mining eventualities. To the best of our data, we are the primary to explore this new insider threat in DKBDM. We have a tendency to conjointly analytically demonstrate the minimum amount of insider information necessary to launch the insider attack. Finally, we have a tendency to follow up by introducing many proposed privacy-preserving schemes to counter the described attack.