核心概念
Developing a novel rumor mitigation paradigm using deep reinforcement learning to counter misinformation on social media platforms.
摘要
Social media platforms face challenges from the rapid spread of rumors, impacting public safety and democracy. Existing approaches like suspending users or broadcasting real information are costly and disruptive. A new approach is introduced, minimizing user disturbance by intervening in the social network to slow rumor propagation. A knowledge-informed agent is developed using graph neural networks and policy networks for link selection. Experiments show over 25% reduction in affected populations. The proposed method is released as open-source code.
統計資料
Experiments demonstrate over 25% reduction in affected populations.
Removal of 10% of influencing relationships while retaining at least 60% for each user.
Improvement compared to baselines ranges from 19% to 25%.
引述
"We propose a knowledge-informed agent trained with a randomized algorithm, which blocks a minimal set of links to mitigate rumors."
"Our method demonstrates substantial effectiveness in curbing the spread of misinformation, showcasing an improvement of over 21.1% compared to baseline methods."