Heuristic Gradient Optimization Approach to Controlling Susceptibility to Manipulation in Online Social Networks
Document Type
Conference Proceeding
Publication Date
1-1-2023
Abstract
Manipulation through inferential attacks in online social networks (OSN) can be achieved by learning the user’s interests through their network and their interactions with the network. Since some users have a higher propensity for disclosure than others, a one-size-fits-all technique for limiting manipulation proves insufficient. In this work, we propose a model that allows the user to adjust their online persona to limit their susceptibility to manipulation based on their preferred disclosure threshold. Our experiment, using real-world data provides a way to measure manipulation gained from a single tweet. We then proffer solutions that show that manipulation gain derived as a result of participating in OSNs can be minimized and adjusted to meet the user’s needs and expectations, giving at least some measure of control to the user.
Publication Source (Journal or Book title)
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
First Page
167
Last Page
178
Recommended Citation
Osho, A., Wei, S., & Amariucai, G. (2023). Heuristic Gradient Optimization Approach to Controlling Susceptibility to Manipulation in Online Social Networks. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 13831 LNCS, 167-178. https://doi.org/10.1007/978-3-031-26303-3_15
