[Paper Review] Asynchronous Opinion Dynamics in Social Networks

Petra Berenbrink, Martin Hoefer+4 more|arXiv (Cornell University)|Jan 30, 2022

Opinion Dynamics and Social Influence4 citations

TL;DR

This paper analyzes asynchronous opinion dynamics in social networks using the Hegselmann-Krause model, where agents update their opinions based on neighbors with similar views. It proves convergence to a δ-stable state in O(n|E|²(ε/δ)²) expected updates for any social network, significantly improving prior bounds and showing tightness via simulations on path and dumbbell graphs.

ABSTRACT

Opinion spreading in a society decides the fate of elections, the success of products, and the impact of political or social movements. The model by Hegselmann and Krause is a well-known theoretical model to study such opinion formation processes in social networks. In contrast to many other theoretical models, it does not converge towards a situation where all agents agree on the same opinion. Instead, it assumes that people find an opinion reasonable if and only if it is close to their own. The system converges towards a stable situation where agents sharing the same opinion form a cluster, and agents in different clusters do not \mbox{influence each other.} We focus on the social variant of the Hegselmann-Krause model where agents are connected by a social network and their opinions evolve in an iterative process. When activated, an agent adopts the average of the opinions of its neighbors having a similar opinion. By this, the set of influencing neighbors of an agent may change over time. To the best of our knowledge, social Hegselmann-Krause systems with asynchronous opinion updates have only been studied with the complete graph as social network. We show that such opinion dynamics with random agent activation are guaranteed to converge for any social network. We provide an upper bound of $\mathcal{O}(n|E|^2 (\varepsilon/δ)^2)$ on the expected number of opinion updates until convergence, where $|E|$ is the number of edges of the social network. For the complete social network we show a bound of $\mathcal{O}(n^3(n^2 + (\varepsilon/δ)^2))$ that represents a major improvement over the previously best upper bound of $\mathcal{O}(n^9 (\varepsilon/δ)^2)$. Our bounds are complemented by simulations that indicate asymptotically matching lower bounds.

Motivation & Objective

• To analyze the convergence behavior of asynchronous opinion dynamics in social networks under the Hegselmann-Krause model.
• To establish provably tight upper bounds on convergence time for arbitrary social network topologies.
• To provide the first non-trivial theoretical lower bound for asynchronous opinion updates in this setting.
• To validate theoretical findings through agent-based simulations on structured networks like paths and dumbbell graphs.
• To extend prior work limited to complete graphs by analyzing general network structures.

Proposed method

• The authors model opinion dynamics as a potential function-based process, where each agent updates its opinion to the average of neighbors within a confidence bound ε.
• They use a potential function Φ(S) = 1/2 ∑_{i,j} w_ij (x_i - x_j)² to track system progress, with w_ij = 1 if |x_i - x_j| ≤ ε and 0 otherwise.
• The analysis bounds the expected decrease in potential per update, leveraging random agent activation and the δ-stability condition.
• They derive an upper bound of O(n|E|²(ε/δ)²) on the expected number of updates until δ-stability is reached.
• Simulations on path and dumbbell graphs validate the theoretical bounds, showing convergence times scaling as Θ(n³) and Θ(n⁴), respectively.
• The method combines stochastic analysis with structural properties of the social network to bound convergence speed.

Experimental results

Research questions

• RQ1 Does asynchronous opinion dynamics in the Hegselmann-Krause model converge to a δ-stable state for any arbitrary social network?
• RQ2 What is the expected convergence time of asynchronous updates in general social networks, and how does it scale with network size and structure?
• RQ3 Can the theoretical upper bound on convergence time be shown to be asymptotically tight for specific network topologies?
• RQ4 How does the convergence behavior differ between structured networks like paths and more complex topologies like dumbbell graphs?
• RQ5 What is the first non-trivial lower bound for convergence time in the asynchronous update model?

Key findings

• The paper establishes an upper bound of O(n|E|²(ε/δ)²) on the expected number of opinion updates until convergence to a δ-stable state for any social network.
• For the complete graph, the bound improves to O(n³(n² + (ε/δ)²)), representing a major improvement over the prior O(n⁹(ε/δ)²) bound.
• Simulations on path graphs show convergence time scaling as Θ(n³), matching the theoretical upper bound.
• Simulations on dumbbell graphs indicate a convergence time scaling as Θ(n⁴), suggesting the upper bound is tight in this case.
• The study provides the first proven non-trivial lower bound for asynchronous opinion dynamics, supporting the tightness of the theoretical upper bound.