Explainability and Graph Learning From Social Interactions

Shumovskaia, Valentina; Ntemos, Konstantinos; Vlaski, Stefan; Sayed, Ali H.

doi:10.1109/tsipn.2022.3223805

Cited by 7 publications

(9 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The following boundedness assumption on the likelihood is common in the literature [7], [27]; it essentially amounts to assuming that the likelihoods share support regions.…”

Section: Social Learning Modelmentioning

confidence: 99%

“…It was shown in [6], [27] that this social learning scheme has a powerful performance guarantee. Specifically, the probability of error goes to zero for both the private and public beliefs as the step-size δ approaches zero, namely,…”

Section: Assumption 1 (Bounded Likelihoods)mentioning

confidence: 99%

“…In other words, the agents are able to learn the truth eventually. Following [27], we introduce two matrices Λ i and L i in order to represent the recursions (3)-(4) in a more compact matrix form as follows:…”

Section: Assumption 1 (Bounded Likelihoods)mentioning

confidence: 99%

“…We have performed some prior work on explainability for social learning [27], [28]. In these works, given the evolution of beliefs and assuming some partial prior knowledge about the distribution of private observations, it was shown how to identify pairwise influences inside the network (i.e., how strongly pairs of agents influence each other), as well as how to recover the underlying graph topology.…”

Section: Introduction and Related Workmentioning

confidence: 99%

“…There have also been several important contributions to graph learning such as [27], [28], [50]- [63]. This is because the graph structure plays an important role in distributed learning [64]- [67], and its identification can provide valuable information about relationships within the network [27]. One approach is based on exploiting correlations and similarities between vertices [68]- [70].…”

Section: Introduction and Related Workmentioning

confidence: 99%

See 4 more Smart Citations

Discovering Influencers in Opinion Formation Over Social Graphs

Shumovskaia

Kayaalp

Cemri

et al. 2023

IEEE Open J. Signal Process.

Self Cite

View full text Add to dashboard Cite

The adaptive social learning paradigm helps model how networked agents are able to form opinions on a state of nature and track its drifts in a changing environment. In this framework, the agents repeatedly update their beliefs based on private observations and exchange the beliefs with their neighbors. In this work, it is shown how the sequence of publicly exchanged beliefs over time allows users to discover rich information about the underlying network topology and about the flow of information over the graph. In particular, it is shown that it is possible (i) to identify the influence of each individual agent to the objective of truth learning, (ii) to discover how well-informed each agent is, (iii) to quantify the pairwise influences between agents, and (iv) to learn the underlying network topology. The algorithm derived herein is also able to work under non-stationary environments where either the true state of nature or the graph topology are allowed to drift over time. We apply the proposed algorithm to different subnetworks of Twitter users, and identify the most influential and central agents by using their public tweets (posts).

show abstract

“…The following boundedness assumption on the likelihood is common in the literature [7], [27]; it essentially amounts to assuming that the likelihoods share support regions.…”

Section: Social Learning Modelmentioning

confidence: 99%

Section: Assumption 1 (Bounded Likelihoods)mentioning

confidence: 99%

Section: Assumption 1 (Bounded Likelihoods)mentioning

confidence: 99%

Section: Introduction and Related Workmentioning

confidence: 99%