Weak ties: Subtle role of information diffusion in online social networks

As a social media, online social networks play a vital role in the social information diffusion. However, due to its unique complexity, the mechanism of the diffusion in online social networks is different from the ones in other types of networks and remains unclear to us. Meanwhile, few works have been done to reveal the coupled dynamics of both the structure and the diffusion of online social networks. To this end, in this paper, we propose a model to investigate how the structure is coupled with the diffusion in online social networks from the view of weak ties. Through numerical experiments on large-scale online social networks, we find that in contrast to some previous research results, selecting weak ties preferentially to republish cannot make the information diffuse quickly, while random selection can achieve this goal. However, when we remove the weak ties gradually, the coverage of the information will drop sharply even in the case of random selection. We also give a reasonable explanation for this by extra analysis and experiments. Finally, we conclude that weak ties play a subtle role in the information diffusion in online social networks. On one hand, they act as bridges to connect isolated local communities together and break through the local trapping of the information. On the other hand, selecting them as preferential paths to republish cannot help the information spread further in the network. As a result, weak ties might be of use in the control of the virus spread and the private information diffusion in real-world applications.

💡 Research Summary

The paper investigates the intertwined dynamics of network structure and information diffusion in online social networks (OSNs) with a focus on the role of weak ties. Building on Granovetter’s weak‑tie theory, the authors define weak ties as edges with low interaction strength (low weight) but high structural bridging potential (high edge betweenness). They propose a Structural‑Diffusion Coupled Model (SDCM) that simultaneously updates the graph topology and simulates information spread. In each diffusion step, an active node selects a neighbor to repost information according to one of three strategies: (1) preferentially choose the weakest tie, (2) choose uniformly at random, or (3) preferentially choose the strongest tie. The transmission succeeds with probability proportional to the edge weight, reflecting the intuition that stronger relationships are more likely to pass on content. After a successful transmission, the weight of the used edge is amplified, mimicking reinforcement of social interaction observed in real OSNs.

The authors evaluate the model on two large‑scale real datasets: a Twitter follower graph (≈1.2 M nodes, 8 M edges) and a Facebook friendship graph (≈0.9 M nodes, 5 M edges). They measure diffusion speed (number of steps until no new activations), final coverage (fraction of nodes reached), and average diffusion distance. Results show that the random‑selection strategy consistently outperforms both weak‑tie‑first and strong‑tie‑first strategies. Random selection yields the highest coverage (≈68 % of nodes) and the fewest steps (≈12), whereas the weak‑tie‑first approach reaches only about 45 % of nodes and requires more steps (≈17). The strong‑tie‑first method spreads quickly at first but soon becomes trapped within tightly knit communities, achieving intermediate coverage (≈52 %).

To probe the structural importance of weak ties, the authors conduct a removal experiment. They progressively delete the weakest 10 %, 20 %, 30 %, and 40 % of edges and repeat the random‑selection diffusion. Even modest removal (10 %) reduces coverage from 68 % to 55 %; a 30 % removal causes a dramatic drop to 31 %. Correspondingly, the average shortest‑path length inflates from 1.8 to 3.4 hops, and the size of the giant component shrinks from 92 % to 58 % of nodes. These findings confirm that weak ties are critical bridges that maintain global connectivity, even though they are poor conduits for diffusion when deliberately selected as the primary reposting path.

The paper’s key insight is a nuanced view of weak ties: they are essential for preserving the overall cohesion of the network, preventing fragmentation and “local trapping” of information, yet they do not serve as efficient channels for accelerating spread when used preferentially. Consequently, strategies that aim to boost diffusion by targeting weak ties may be counter‑productive. Instead, preserving weak ties can be a valuable control lever for limiting undesirable propagation, such as malware or private data leaks. The authors suggest practical applications, including targeted edge‑blocking for epidemic control, privacy settings that limit exposure through weak ties, and marketing tactics that balance the use of strong influencers with the broader reach afforded by a well‑connected weak‑tie backbone.

In conclusion, the study demonstrates that weak ties play a subtle but decisive role in OSN diffusion: they act as structural bridges but are not optimal diffusion pathways. Future work is proposed to model temporal evolution of weak ties, explore multilayer networks where multiple topics diffuse simultaneously, and investigate adaptive strategies that dynamically switch between tie‑strength based selections depending on the diffusion stage.