Source-linked AI summary
Small world yields the most effective information spreading
Linyuan Lü, Duan-Bing Chen, Tao Zhou
TL;DR
The paper addresses the gap between epidemic and information-spreading models by proposing a network model with information-specific mechanisms. It compares network structures and sizes, finding that regular networks can outperform random ones under some conditions, while slight randomness can maximize spreading effectiveness.
Problem
Prior work often used unified epidemic and information-spreading models, despite information involving memory, social reinforcement, and judged authenticity.
Method
The paper proposes an SIR variant with four news states, cumulative-exposure approval, three information-specific rules, and comparisons across degree-preserving network structures.
Results
Regular networks spread information faster and broader than random networks at low λ, whereas small-world networks can achieve the most effective spreading when social reinforcement is strong.
Takeaways & Limitations
Information-spreading effectiveness depends on local clustering, randomness, social reinforcement, and network size rather than following epidemic-spreading patterns universally.
Abstract
from arXiv · showhide
Spreading dynamics of information and diseases are usually analyzed by using a unified framework and analogous models. In this paper, we propose a model to emphasize the essential difference between information spreading and epidemic spreading, where the memory effects, the social reinforcement and the non-redundancy of contacts are taken into account. Under certain conditions, the information spreads faster and broader in regular networks than in random networks, which to some extent supports the recent experimental observation of spreading in online society [D. Centola, Science {\bf 329}, 1194 (2010)]. At the same time, simulation result indicates that the random networks tend to be favorable for effective spreading when the network size increases. This challenges the validity of the above-mentioned experiment for large-scale systems. More significantly, we show that the spreading effectiveness can be sharply enhanced by introducing a little randomness into the regular structure, namely the small-world networks yield the most effective information spreading. Our work provides insights to the understanding of the role of local clustering in information spreading.
I. INTRODUCTION
The paper distinguishes information spreading from epidemic spreading and proposes a model incorporating information-specific mechanisms. It predicts that network structure and size determine which topology spreads information most effectively.
- Information spreading includes opinions and rumors whose value or authenticity individuals must judge, unlike infectious diseases requiring physical contact.
- Earlier unified models emphasized similarities between epidemic and information spreading while overlooking their essential differences.
- The proposed SIR variant incorporates memory effects, social reinforcement, and non-redundancy of contacts.
- For spreading rates below λ∗, regular networks can spread information more effectively than random networks, supporting Centola’s small-network observation.
- Small-world networks yield the most effective information spreading when a little randomness is introduced into regular structure.
II. MODEL
The model represents news diffusion through four individual states and approval decisions shaped by cumulative exposure. It compares ordered, rewired, and random networks while preserving node degree.
- Individuals occupy Unknown, Known, Approved, or Exhausted states, distinguishing awareness, approval, transmission, and post-transmission inactivity.
- A randomly selected seed starts the process by transmitting the news to all neighbors before becoming exhausted.
- Approval depends on cumulative receptions through P(m) = (λ − T)e^−b(m−1) + T, with λ as first-receipt probability and T as its upper bound.
- The parameter b controls social reinforcement: P(m) approaches T as cumulative exposure m increases, with larger b indicating stronger reinforcement.
- The study compares ordered regular, homogeneous small-world, and homogeneous random networks while maintaining identical node degree k.
III. RESULTS
Information spreading depends strongly on network structure and social reinforcement: regular networks can outperform random ones at low spreading rates, while small-world networks are most effective when reinforcement is substantial.
- Regular versus random networks: At low λ, spreading is faster and broader on regular networks than on random networks, but random networks become favorable after λc ≈0.145.The crossing point marks where the final number of approved nodes on random networks exceeds that on regular networks.
- Regular versus random networks: As network size increases, λc decreases and then becomes insensitive to N, weakening the regular-network advantage in large systems.The authors therefore suggest that Centola’s experimental result may not hold, or may be weakened, at large scale.
- Small-world effectiveness: A tiny randomness p = 0.02 raises R from 205 on regular networks to 6593, exceeding the random-network value R = 4049 when b = 0.8.The final number of approved nodes is non-monotonic in randomness, with an optimal p∗ rather than continual improvement.
- Small-world effectiveness: The results indicate that local clustering can enhance the approving rate of information, contrasting with its role in traditional epidemic-spreading results.This conclusion connects the network-structure findings to the paper’s distinction between information and epidemic spreading.
- Small-world effectiveness: Stronger social reinforcement produces smaller optimal randomness p∗, shifting the most effective structure toward small-world networks.With weak reinforcement, p∗ is close to 1 and the ordering is Random > Small-World > Regular; with strong reinforcement, small-world networks are most effective.
- Small-world effectiveness: For N = 10000, small-world networks have p(R > 10) = 0.703 versus 0.460 for random networks, and can reach 9900 individuals with probability 0.684.The maximum reach reported for regular networks is only 1680.
IV. CONCLUSION AND DISCUSSION
The paper argues that information spreading should not be treated as equivalent to epidemic spreading. It calls for studying their essential differences in large-scale online systems, where detailed analysis is increasingly feasible.
- Discussion: Information spreading differs from epidemic spreading in ways that motivate models focused on information-specific dynamics rather than a unified framework.The authors identify time-decaying effects as one significant difference and argue that prior studies have overemphasized the two processes’ similarity.
- Discussion: Large-scale online systems now make detailed analysis of information spreading feasible through advances in database technology and computational power.The paper presents this feasibility as a reason to investigate information spreading directly.