Abstract:Social networks have attracted billions of users and supported a wide range of interests and practices. Users of social networks can be connected with each other by different communities according to professions, living locations, and personal interests. With the development of diverse social network applications, academic researchers, and practicing engineers pay increasing attention to the related technology. As each user on the social network platforms typically stores and shares a large amount of personal data, the privacy of such user-related information raises serious concerns. Most research on privacy protection relies on specific information security techniques such as anonymization or access control. However, the protection of privacy depends heavily on the incentive mechanisms of social networks, like users' psychological decisions on security execution and socio-economic considerations. For example, the desire to influence the behaviors of other people may change a user's choice of security setting. In this paper, a game theoretic framework is established to model users' interactions that influence users' decisions as to whether to undertake privacy protection or not. To model the relationship of user communities, community-structured evolutionary dynamics are introduced, in which interactions of users can only happen among those users who have at least one community in common. Then the dynamics of the users' strategies to take a specific privacy protection or not is analyzed based on the proposed community structured evolutionary game theoretic framework. Experiments show that the proposed framework is effective in modeling the users' relationships and privacy protection behaviors. Moreover, results can also help social network managers to design appropriate security service and payment mechanisms to encourage their users to take the privacy protection, which can promote the spreading of privacy behavior throughout the network.

The QoS and Privacy Trade-off of Adversarial Deep Learning: an Evolutionary Game Approach.

Privacy-Preserving Collaborative Deep Learning with Unreliable Participants.

Privacy Risks of Social Interaction Structure: Network Learning in Quadratic Games

A Privacy Protection Model of Data Publication Based on Game Theory

LSTEG: An evolutionary game model leveraging deep reinforcement learning for privacy behavior analysis on social networks

Evolutionary privacy-preserving learning strategies for edge-based IoT data sharing schemes

Community-Structured Evolutionary Game for Privacy Protection in Social Networks.

An Incentive Mechanism for Privacy-Preserving Crowdsensing via Deep Reinforcement Learning

Privacy Protection: A Community-Structured Evolutionary Game Approach

Differentially-Private Deep Learning from an Optimization Perspective.

QoS Prediction and Adversarial Attack Protection for Distributed Services under DLaaS

Evolutionary Game Based Analysis for User Privacy Protection Behaviors in Social Networks.

Private Deep Neural Network Models Publishing for Machine Learning As a Service.

An Evolutionary Game-Theoretic Approach for Assessing Privacy Protection in Mhealth Systems

QI-DPFL: Quality-Aware and Incentive-Boosted Federated Learning with Differential Privacy

An Evolutionary Game Theory-Based Cooperation Framework for Countering Privacy Inference Attacks

How to dissolve the "privacy paradox" in social networks? A game approach based on privacy calculus

Privacy-Aware Data Trading

Differential Private Collaborative Web Services QoS Prediction

Dynamic Games for Social Model Training Service Market via Federated Learning Approach

A QoS Evolutionary Method of Cloud Service Based on User Utility Model