Polarized information ecosystems can reorganize social networks via information cascades

The precise mechanisms by which the information ecosystem polarizes society remain elusive. Focusing on political sorting in networks, we develop a computational model that examines how social network structure changes when individuals participate in information cascades, evaluate their behavior, and potentially rewire their connections to others as a result. Individuals follow proattitudinal information sources but are more likely to first hear and react to news shared by their social ties and only later evaluate these reactions by direct reference to the coverage of their preferred source. Reactions to news spread through the network via a complex contagion. Following a cascade, individuals who determine that their participation was driven by a subjectively “unimportant” story adjust their social ties to avoid being misled in the future. In our model, this dynamic leads social networks to politically sort when news outlets differentially report on the same topic, even when individuals do not know others’ political identities. Observational follow network data collected on Twitter support this prediction: We find that individuals in more polarized information ecosystems lose cross-ideology social ties at a rate that is higher than predicted by chance. Importantly, our model reveals that these emergent polarized networks are less efficient at diffusing information: Individuals avoid what they believe to be “unimportant” news at the expense of missing out on subjectively “important” news far more frequently. This suggests that “echo chambers”—to the extent that they exist—may not echo so much as silence.

Deconstructing scale-free neuronal avalanches: behavioral transitions and neuronal response

Observations of neurons in a resting brain and neurons in cultures often display spontaneous scale-free collective dynamics in the form of information cascades, also called “neuronal avalanches”. This has motivated the so called critical brain hypothesis which posits that the brain is self-tuned to a critical point or regime, separating exponentially-growing dynamics from quiescent states, to achieve optimality. Yet, how such optimality of information transmission is related to behaviour and whether it persists under behavioural transitions has remained a fundamental knowledge gap. Here, we aim to tackle this challenge by studying behavioural transitions in mice using two-photon calcium imaging of the retrosplenial cortex -- an area of the brain well positioned to integrate sensory, mnemonic, and cognitive information by virtue of its strong connectivity with the hippocampus, medial prefrontal cortex, and primary sensory cortices. Our work shows that the response of the underlying neural population to behavioural transitions can vary significantly between different sub-populations such that one needs to take the structural and functional network properties of these sub-populations into account to understand the properties at the total population level. Specifically, we show that the retrosplenial cortex contains at least one sub-population capable of switching between two different scale-free regimes, indicating an intricate relationship between behaviour and the optimality of neuronal response at the subgroup level. This asks for a potential reinterpretation of the emergence of self-organized criticality in neuronal systems.

Politicization of discussions of Covid-19 on Twitter

Power of social media including Twitter for English speaking community to shape public opinion becomes critical during the current pandemic because of misinformation. The existing studies on spreading misinformation on social media hypothesise that the initial message is fake. In contrast, we focus on information distortion occurring in cascades as the initial message about the Covid-19 treatment is quoted or receives a reply. Public persons discuss medical information on Twitter providing fast and simple response to complex medical problems that users find very attractive to follow. Followers generate information cascades while quoting and commenting on the initial message. In the cascades, medical information from the initial tweet is often distorted. The discussion of the Covid-19 treatment in the cascades is politicized according to users’ political sympathies. We show a significant information shift in cascades initiated by public figures during the Covid-19 pandemic. The study provide valuable insights for the semantic analysis of information distortion.

Using Response Times to Infer Others’ Private Information: An Application to Information Cascades

The standard assumption in social learning environments is that agents learn from others through choice outcomes. We argue that in many settings, agents can also infer information from others’ response times (RT), which can increase efficiency. To investigate this, we conduct a standard information cascade experiment and find that RTs do contain information that is not revealed by choice outcomes alone. When RTs are observable, subjects extract this private information and are more likely to break from incorrect cascades. Our results suggest that in environments where RTs are publicly available, the information structure may be richer than previously thought. This paper was accepted by Yan Chen, decision analysis.

Information Cascades Prediction With Graph Attention

The cascades prediction aims to predict the possible information diffusion path in the future based on cascades of the social network. Recently, the existing researches based on deep learning have achieved remarkable results, which indicates the great potential to support cascade prediction task. However, most prior arts only considered either cascade features or user relationship network to predict cascade, which leads to the performance limitation because of the lack of unified modeling for the potential relationship between them. To that end, in this paper, we propose a recurrent neural network model with graph attention mechanism, which constructs a seq2seq framework to learn the spatial-temporal cascade features. Specifically, for user spatial feature, we learn potential relationship among users based on social network through graph attention network. Then, for temporal feature, a recurrent neural network is built to learn their structural context in several different time intervals based on timestamp with a time-decay attention. Finally, we predict the next user with the latest cascade representation which obtained by above method. Experimental results on two real-world datasets show that our model achieves better performance than the baselines on the both evaluation metrics of HITS and mean average precision.

Opinion Diversity and the Resilience of Cooperation in Dynamical Networks

Across various scenarios, individuals cooperate with others to contribute towards a shared goal and ensure self-preservation. In game theory, the act of cooperation is considered as an individual producing some form of benefit to be utilised by others, under the expectation others will return the favour. In several scenarios, individuals make use of their own information to aid with their decision about who to connect and cooperate with. However, the choice of cooperation can be taken advantage of by opportunistic defectors, which can lead to significant disruption. This paper investigates how the diversity of opinion can contribute to the structure and mechanics of a dynamical network model and to the resilience of cooperation, by utilising a computational model where individuals make use of both public and private information to implement their decision. Our results show that increasing diversity leads to more stable, less connected and less prosperous networks coupled to more frequent, but shallower information cascades. Our work generally shows that the outcome of the conflict between cooperators and cheaters strongly depends on the interplay between population structure, individual decision making and individual opinions.