The online social network has been widely used among the population. It is no surprise that individuals are more likely to get affected by their social contacts while adopting new ideas. This motives parities or companies with comparable opinions and products to identify "opinion leaders" and market via the "word-of-mouth" effect.
In this project, we are proposing a framework Multidimensional Uncertainty-aware Deep Reinforcement learning-based competitive Influence Maximization (muDRIM) to help combat false information spreading while maximizing the Influence of True information. We employ Deep Reinforcement Learning (DRL) as a decision-making algorithm for multiple parties to compete for their influence in the Competitive Influence Maximization (CIM) problems over a partially observed network. We also adopt users' multidimensional uncertain and subjective opinions to reflect a more realistic online opinion competition.
