Opinion polarisation and echo chambers in online social media: structure, dynamics and mitigation strategies

在一个联系越来越紧密的世界里，人际关系和观点都是通过在线互动形成和发展的。想法相似的用户往往聚集在一起，形成强烈固执己见的在线社区，即所谓的回音室。随之而来的意见两极分化助长了阴谋论的传播，阻碍了卓有成效的政治辩论。意见动态的研究越来越关注这一问题，大多数模型都考虑了用户层面的心理和社会特征。然而，这些在实践中难以评估，并且缺乏对这些理论作品的实证评估。在这个博士研究项目中，我将提出一个新的数学框架，用于描述回声室和从宏观角度量化极化。我还将在现实世界的数据上评估我的模型，并为旨在减轻意见两极分化的算法方法提出建议。

In a more and more connected world, relationships and opinions are shaped and evolve through online interactions. Similar-minded users tend to cluster together and form strongly opinionated online communities, the so-called echo chambers. The ensuing polarisation of opinions fosters the spread of conspiracy theories and hinder fruitful political debate. Research on opinion dynamics is focusing more and more on this problem, with most models considering user-level psychological and social features. Those are however difficult to evaluate in practice and there has been a dearth of empirical evaluation of these theoretical works. In this PhD research project I will present a new mathematical framework for the description of echo chambers and quantification of polarisation from a macroscopic point of view. I will also evaluate my model on real-world data and propose leads for algorithmic methods aimed at mitigating the polarisation of opinions.