摘要：Privacy leakage remains a critical challenge in online social networks (OSNs), where users continuously share information despite potential risks. The interconnected nature of OSNs makes the control of privacy information diffusion more complex. This paper proposes a novel adaptive privacy protection framework that can dynamically mitigate the privacy risks during privacy information diffusion. We model privacy diffusion and inference risks along propagation paths and develop a privacy policy generation mechanism using deep reinforcement learning (DRL) to optimize the trade-off between privacy protection and information sharing utility. Extensive experiments on real and synthetic social network datasets show that our method outperforms baseline approaches in controlling privacy risk while preserving the information sharing utility. Moreover, our study highlights the effectiveness of future reward prediction in reinforcement learning for privacypreserving decision-making. ? 2025 IEEE.

ISSN号：2836-3876

卷、期、页：期2025：207-217

发表日期：2025-01-01

期刊分区（SCI为中科院分区）：无

收录情况：EI（工程索引）

发表期刊名称：Proceedings of the IEEE International Conference on Web Services, ICWS

参与作者：罗叶红,何泾沙

第一作者：宜裕紫,吕仲琪,黄霁崴

论文类型：会议论文

论文概要：宜裕紫,罗叶红,何泾沙,吕仲琪,黄霁崴，Towards Adaptive Privacy-Preserving Information Diffusion in Online Social Networks，Proceedings of the IEEE International Conference on Web Services, ICWS，2025，期2025：207-217

论文题目：Towards Adaptive Privacy-Preserving Information Diffusion in Online Social Networks