Abstract: Multi-agent models often describe populations segregated either in the physical space, i.e. subdivided in metapopulations, or in the ecology of opinions, i.e. partitioned in echo chambers. Here we show how the interplay between homophily and social influence controls the emergence of both kinds of segregation in a simple model of mobile agents, endowed with a continuous opinion variable. In the model, physical proximity determines a progressive convergence of opinions but differing opinions result in agents moving away from each others. This feedback between mobility and social dynamics determines to the onset of a stable dynamical metapopulation scenario where physically separated groups of like-minded individuals interact with each other through the exchange of agents. The further introduction of confirmation bias in social interactions, defined as the tendency of an individual to favor opinions that match his own, leads to the emergence of echo chambers where different opinions can coexist also within the same group. We believe that the model may be of interest to researchers investigating the origin of segregation in the offline and online world.

Abstract: In the world of artistic production there is a constant struggle to achieve fame and popularity. This fierce competition between artistic creations results in the emergence of highly popular elements that are usually well remembered throughout the years, while many other works that did not achieve that status are long-forgotten. However, there is another level of importance that must be considered in order to have a more complete picture of the system. In fact many works that have influenced the production itself, both due to their aesthetic and cultural value, might have not been or might not be popular anymore. Due to their relevance for the whole artistic production, it is important to identify them and save their memory for obvious cultural reasons. In this paper we focus on the duality between popularity and significance in the context of popular music, trying to understand the features of music albums belonging to one or both of these classes. By means of user-generated data gathered on Last.fm, an on-line catalog of music albums, we define a growing conceptual space in the form of a network of tags representing the evolution of music production during the years. We use this network in order to define a set of general metrics, characterizing the features of the albums and their impact on the global music production. We then use these metrics to implement an automated prediction method of both the commercial success of a creation and its belonging to expert-made lists of particularly significant and important works. We show that our metrics are not only useful to asses such predictions, but can also highlight important differences between culturally relevant and simply popular products. Finally, our method can be easily extended to other areas of artworks creation.

Abstract: This paper presents a computer model that describes between-agent social interactions as dynamic couplings between limit-cycle oscillators, leading to emergent properties within social networks. The model’s purpose is to represent collective social dynamics as they emerge from local-level interactions, with a specific focus on social rank and status distinctions. This investigation is informed by anthropological observations that societies conventionally construct rank-order differences between individuals. For example, rituals often mark transition between roles, while linguistic conventions – such as Korean grammatical honorifics – enforce differing treatment of people of distinct relative statuses. Such observations raise a puzzling question: why would human societies ritually exaggerate cleavages in their social structure, when unity is typically a desideratum for social collectives? An interdisciplinary complex-systems perspective may be particularly useful here. Psychology studies demonstrate that humans tend to mimic one another automatically during interpersonal interactions. However, hierarchy alters this response, changing it from postural mimicry to complementarity. While somatic and speech mimicry enhances social affiliation, excessive coherence between individuals may also produce instability – as when a fad sweeps across an adolescent peer group. We suggest that, with segmented roles as dampeners, mutual amplification of such behaviors is contained to respective peer subgroups within larger populations. This is an agent-based model, meaning that social actors are simulated using decisional algorithms. Our algorithms are derived from those describing coupled limit-cycle oscillators. Coupling coefficients represent affiliation and mimetic tendencies. Mimesis is suppressed between agents of different social ranks, whereas peers tend to couple readily. Thus, same-rank agents mutually amplify one another’s states across the peer network. Rank distinctions dampen and constrain the propagation of these synchronic states, leading to the emergence of a dynamically structured population. We thereby emphasize a uniquely somatic perspective on the local dynamics that drive emergent social complexity.

Abstract: Empirical evidence shows that in our networked society people have a marked tendency to find themselves surrounded by others who are similar to them, e.g., in language, socio-economic status, educational level, political beliefs, work norm and many others. Two possible causes are: sorting, i.e. people selecting people like them (as in the Schelling's model), and peer effects, i.e. people influenced by people around them (as in the Standing Ovation model). We investigate the dynamics of sorting and peer effects in reaching (or not) mutual coexistence of conventions in a multiplex network topology. We model a social environment through a two-layer multiplex network, in which agents have profiles: a type, i.e. being orange or blue, and a strategy, i.e. rewiring their links or not. The layers can be interpreted as the presence of an informal context, e.g. family, school district, which are mostly fixed over time, and a formal one, e.g. work partnership, in which different agents adaptively restructure their neighbourhood over time. Each agent has a tolerance threshold describing his endurance to a certain level of diversity in its neighbourhood before switching type or strategy. Consequently, agents act according to a mixed motive imitation across the two layers: they conform their strategy to the most frequent one in their neighbourhood on the informal layer and they implement such strategy to their links on the formal, adaptive, layer. Strategies and types are randomly distributed in the beginning on two random graph layered topologies. We observe that the initial fraction of rewirers drives multiple and stable final configurations, going from coexistence to polarization of types, through a tipping point. Secondly, the lower the tolerance the more likely segregation takes place. Finally, we relate those results to either choice or opportunity-based homophily in the system.

Abstract: In the last two decades, Evolutionary Game Theory (EGT) has strongly developed into a mature field that, nowadays, represents a vivid and independent research area with a list of applications. The Public Goods Game (PGG hereinafter) represents the typical game-theoretical framework in which individual and group interests collide. In fact, in the PGG game players can decide to mutually cooperate for achieving a common goal and, at the same time, are tempted to exploit their opponents in order to obtain an higher payoff. A series of works unveiled that even when agents' interactions are based upon Nash Equilibria in which 'defection' theoretically dominates cooperative strategies, agent, and even human, populations are able to attain a cooperative equilibrium. As result, it is interesting to identify behaviors and properties that may lead a population towards cooperation. It is also worth observing that conformism is one of the most investigated behaviors in the field of sociophysics. In the present work ([1]), we investigate the spatial PGG by considering a population composed of conformity-driven agents and fitness-driven agents. Thus, while the former tend to update their strategy with the most adopted one in their neighborhood, the latter tend to imitate their richest neighbor (i.e., the most fitted). Results show that conformism has a prominent role in the spatial PGG: it seems that this social influence may lead the population towards different phases and behaviors, as full cooperation and bistable equilibria. Beyond to present the results of our investigation, we aim to give an brief view of EGT for those participants coming from different communities that can be interested in evolutionary dynamics. [1] JAVARONE MA, ANTONIONI A., CARAVELLI F.: Conformity-driven agents support ordered phases in the spatial public goods game. EPL forthcoming (2016)

Abstract: A key element in archaeological research consists of tracing and understanding change in the past. Emphasis is commonly placed on developments in social, political, and economic organisation. Although the archaeological record is ultimately produced within a societal context, it cannot be interpreted as a clear-cut reflection of this society. The archaeological record is fragmentary and limited as only traces of those activities having a durable material component can be observed. What we see in the archaeological record is not societal organisation as such, but rather the material end-result of social practices. To study past societies we must therefore first understand how practices are bundled to function as constituent elements of societal organisation. The main mechanisms of bundling social practices can be found in the structuration of human action along dimensions of time and space. The former is commonly subdivided in short, medium, and long timespans, the latter in local, regional, and supra-regional scales. Often a convergence is presumed between corresponding scales in time and place on the one hand, and social dynamics on the other. These approaches, however, leave distinct scales in isolation, lacking a proper integration of dynamics across different scales, and remain trapped in a reductionist approach of human societies. To better understand the emergence of social organisation out of the multitude of interactions between human agents, theoretical inspiration has in recent years been sought and found in complex systems thinking. Development of social dynamics and practices in the past can be studied within a resilience-based framework of complex adaptive systems. In this presentation the concept of nested adaptive cycles in particular will be highlighted as a highly potent heuristic tool to move towards an integrated multi-scalar analysis of the temporal and spatial properties of social practices and dynamics, and better understanding of social organisation in past societies.