Studies in social network analysis have identified the effect of echo chambers, i.e. self-reinforcing opinion clusters. Yet, these have been studied on opinions in isolation from beliefs held by participants. We aim at modelling them in close relations with these beliefs.
The process of public opinion formation through interactions between agents has been studied in the research area of opinion dynamics. In the models proposed in this field, agents update their opinions which are usually represented as numerical values through their interactions. Existing opinion dynamics models can be applied in other diciplines [2] or to explain phenomena such as echo chambers in social networks [1].
Recently, a model which considers dynamic topology and timelines in social networks has been proposed [3]. The purpose of the model is to explain why echo chambers (opinion polarization and network segregation in their words) emerge in social networks. Especially, the authors have focused on the bias introduced by algorithms in social networks.
However, these models overlook the effect of what agents know or believe. In most of the cases, we exchange not only our opinions but also our beliefs. The justification of opinions by beliefs can be highlighted in social networks by posts of the form: “Because of (my beliefs in) B, I have opinion O towards the topic φ”. or “My opinion is O towards the topic φ; any other choice would be in contradiction with (my belief) B”. Opinions and beliefs are different: opinions of an agent denote how positive it is toward some topics while beliefs represent what it considers as true.
We aim at studying the emergence and maintenance of echo chambers effect in both beliefs and opinions. For that purpose, we will develop a model with opinions and beliefs in the situation similar to existing social networks. Interactions between opinions can be described by existing opinion dynamics models. Similarly, the process of belief update and propagation can be defined by frameworks proposed in the area of belief revision, which is independent from opinion dynamics. Belief revision game [4] is one of the examples of such frameworks. The interaction between opinions and beliefs will have to be developed based on primary work using values. Using the model, we examine the relations between opinions/beliefs and echo chamber using the model.
The kind of questions that we plan to address are:
The work could be expected to unfold as follows:
It is expected that the experimental design and results be stored in our repository for reproducibility and extendibility purposes.
References:
[1] Faisal Alatawi, Lu Cheng, Anique Tahir, Mansooreh Karami, Bohan Jiang, Tyler Black, Huan Liu. A Survey on Echo Chambers on Social Media: Description, Detection and Mitigation. 2021. arXiv:2112.05084.
[2] Yucheng Dong, Min Zhan, Gang Kou, Zhaogang Ding, Haiming Liang. “A survey on the fusion process in opinion dynamics”. Information fusion 43:57–65 (2018). issn: 1566-2535. doi:10.1016/j.inffus.2017.11.009.
[3] Kazutoshi Sasahara, Wen Chen, Hao Peng, Giovanni Luca Ciampaglia, Alessandro Flammini, Filippo Menczer. “Social influence and unfollowing accelerate the emergence of echo chambers”. Journal of computational social science 4(1):381–402 (Sept. 2020). issn: 2432-2725. doi:10.1007/s42001-020-00084-7.
[4] Nicolas Schwind, Katsumi Inoue, Gauvain Bourgne, Sébastien Konieczny, Pierre Marquis. “Belief Revision Games”. In: Proceedings of the 29th AAAI Conference, Austin (TX US). 2015, pp. 1590–1596. doi:10.1609/aaai.v29i1.9415.
Links:
Reference number: Proposal n°3137
Master profile: M2, research oriented.
Advisor: Jérôme Euzenat (Jerome:Euzenat#inria:fr) and Koji Hasebe (hasebe#cs:tsukuba:ac:jp)
Team: The work will be carried out in the mOeX team common to INRIA & Université Grenoble Alpes. mOeX is dedicated to study knowledge evolution through adaptation. It gather permanent researchers from the Exmo team which has taken an active part these past 15 years in the development of the semantic web.
Laboratory: LIG.
Place of work: The position is located at INRIA Grenoble Rhône-Alpes, Montbonnot (near Grenoble, France) a main computer science research lab, in a stimulating research environment.
Perspectives: There is possibility to pursue in PhD, especially related to this topic.
Procedure: Contact us and provide vitæ and possibly motivation letter and references.