Complexity Digest

Stefano Nolfi

Adaptive Behavior

Large Language Models (LLMs) are capable of displaying a wide range of abilities that are not directly connected with the task for which they are trained: predicting the next words of human-written texts. In this article, I review recent research investigating the cognitive abilities developed by LLMs and their relation to human cognition. I discuss the nature of the indirect process that leads to the acquisition of these cognitive abilities, their relation to other indirect processes, and the implications for the acquisition of integrated abilities. Moreover, I propose the factors that enable the development of abilities that are related only very indirectly to the proximal objective of the training task. Finally, I discuss whether the full set of capabilities that LLMs could possibly develop is predictable.

Read the full article at: journals.sagepub.com

DeVerna MR, Aiyappa R, Pacheco D, Bryden J, Menczer F

PLoS ONE 19(5): e0302201

The world’s digital information ecosystem continues to struggle with the spread of misinformation. Prior work has suggested that users who consistently disseminate a disproportionate amount of low-credibility content—so-called superspreaders—are at the center of this problem. We quantitatively confirm this hypothesis and introduce simple metrics to predict the top superspreaders several months into the future. We then conduct a qualitative review to characterize the most prolific superspreaders and analyze their sharing behaviors. Superspreaders include pundits with large followings, low-credibility media outlets, personal accounts affiliated with those media outlets, and a range of influencers. They are primarily political in nature and use more toxic language than the typical user sharing misinformation. We also find concerning evidence that suggests Twitter may be overlooking prominent superspreaders. We hope this work will further public understanding of bad actors and promote steps to mitigate their negative impacts on healthy digital discourse.

Read the full article at: journals.plos.org

Daniel Rico and Yérali Gandica

Entropy 2024, 26(5), 363

Social media has dramatically influenced how individuals and groups express their demands, concerns, and aspirations during social demonstrations. The study of X or Twitter hashtags during those events has revealed the presence of some temporal points characterised by high correlation among their participants. It has also been reported that the connectivity presents a modular-to-nested transition at the point of maximum correlation. The present study aims to determine whether it is possible to characterise this transition using entropic-based tools. Our results show that entropic analysis can effectively find the transition point to the nested structure, allowing researchers to know that the transition occurs without the need for a network representation. The entropic analysis also shows that the modular-to-nested transition is characterised not by the diversity in the number of hashtags users post but by how many hashtags they share.

Read the full article at: www.mdpi.com

SFI Press launched Foundational Papers in Complexity Science this May. The four-volume collection, edited by SFI President David Krakauer, contains 88 typeset articles, each with introductions and commentaries from leading scientists.

Volumes 1 and 2, featuring papers spanning the years 1922 to 1973, are out now. The third and fourth volumes include work from the final decades of the 20th century and will be published later this summer.

The Santa Fe Institute is sponsoring a book set for each of the upcoming Emerging Award Winners, to be announced during the next Conference on Complex Systems in Exeter, UK in September.

More at: cssociety.org

Vivek Jadhav, Roberto Pasqua, Christophe Zanon, Matthieu Roy, Gilles Tredan, Richard Bon, Vishwesha Guttal, Guy Theraulaz

Across taxa, group-living organisms exhibit collective escape responses to stimuli varying from mild stress to predatory pressures. How exactly does information flow among group members leading to a collective escape remains an open question. Here we study the collective responses of a flock of sheep to a shepherd dog in a driving task between well-defined target points. We collected high-resolution spatio-temporal data from 14 sheep and the dog, using Ultra Wide Band tags attached to each individual. Through the time delay analysis of velocity correlations, we identify a hierarchy among sheep in terms of directional influence. Notably, the average spatial position of a sheep along the front-back axis of the group’s velocity strongly correlates with its impact on the collective movement. Our findings demonstrate that, counter-intuitively, directional information on shorter time scales propagates from the front of the group towards the rear, and that the dog exhibits adaptive movement adjustments in response to the flock’s dynamics. Furthermore, we show that a simple shepherding model can capture key features of the collective response of the sheep flocks. In conclusion, our study reveals novel insights on how directional information propagates in escaping animal groups.

Read the full article at: www.biorxiv.org

The recent availability of massive amounts of digital data have profoundly revolutionized research on migration and mobility, enabling scientists to quantitatively study individual and collective mobility patterns at different granularities as generated by human activities in their daily life.

• A satellite event of the Conference on Complex Systems 2024
• Date of Event: September 3, 2024
• Submissions deadline: June 10, 2024

Read the full article at: www.demogr.mpg.de

​”Guided Self-Organization: Machine Learning in Embodied Agents”
The 11th International Conference on Guided Self-Organization takes place during 12-14 February 2025 in Tübingen, Germany. GSO-2025 is organized by The University of Tübingen, The Hamburg University of Technology, The Max Planck Institute for Intelligent Systems, and The International Association for Guided Self-Organization (TIA-GSO).
Research Aims and Topics
The goal of Guided Self-Organization (GSO) is to leverage the strengths of self-organization (simplicity, parallelization, adaptability, robustness, scalability) while still being able to direct the outcome of the self-organizing process. GSO typically has the following features: (i) an increase in organization (structure and/or functionality) over some time; (ii) the local interactions are not explicitly guided by any external agent; (iii) task-independent objectives are combined with task-dependent constraints.

GSO “aims to regulate self-organization for specific purposes, so that a dynamical system may reach specific attractors or outcomes. The regulation constrains a self-organizing process within a complex system by restricting local interactions between the system components, rather than following an explicit control mechanism or a global design blueprint.” Information theory, nonlinear dynamics and network theory are core to many of these methods, and quantifying complexity, its sources and effects is a common theme.
​
The GSO-2025 conference will bring together invited experts and researchers in machine learning, artificial life, self-organizing systems, and complex adaptive systems, with particular emphasis on autonomous agents, information theory, critical phenomena and emergent behaviour. Special topics of interest include: reinforcement learning, intrinsic motivations, origin of life, systems biology, physics of life, unconventional computation, swarm intelligence, measures of complexity, criticality, complex networks, information-driven self-organization (IDSO), etc.

The program includes three days, with five keynote talks, and a number of regular onsite presentations on each day. There are no registration fees for the conference.

More at: www.guided-self.org

Please join us to celebrate 20 years of the Places & Spaces: Mapping Science exhibit! The exhibit is curated here at CNS and has traveled the globe showcasing best examples of information visualization.

When?  June 6, 2024 from 4PM to 6PM EDT.

Where?  In person at University Collections at McCalla, 525 E 9th St., Bloomington, IN 47408 or online via Zoom webinar.

What?  Reception. Enjoy refreshments, remarks from the exhibition curators, presentations from teams whose works have been selected for inclusion in the exhibit this year, and the opportunity to try out a data visualization in VR.

Why?  To introduce the latest additions to the exhibit and celebrate the 20th anniversary of the inception of the Places & Spaces: Mapping Science exhibit, see 100 maps and 40 interactive macroscopes at scimaps.org.

More at: cns-iu.github.io

Andrea Musso and Dirk Helbing

Royal Society Open Science

May 2024 Volume 11Issue 5

Socio-diversity, the variety of human opinions, ideas, behaviours and styles, has profound implications for social systems. While it fuels innovation, productivity and collective intelligence, it can also complicate communication and erode trust. So what mechanisms can influence it? This paper studies how fundamental characteristics of social networks can support or hinder socio-diversity. It employs models of cultural evolution, mathematical analysis and numerical simulations. We find that pronounced inequalities in the distribution of connections obstruct socio-diversity. By contrast, the prevalence of close-knit communities, a scarcity of long-range connections, and a significant tie density tend to promote it. These results open new perspectives for understanding how to change social networks to sustain more socio-diversity and, thereby, societal innovation, collective intelligence and productivity.

Read the full article at: royalsocietypublishing.org

Rafael Prieto-Curiel, Ola Ali, Elma Dervić, Fariba Karimi, Elisa Omodei, Rainer Stütz, Georg Heiler, Yurij Holovatch

PNAS Nexus, Volume 3, Issue 5, May 2024, page 178,

Migration’s impact spans various social dimensions, including demography, sustainability, politics, economy, and gender disparities. Yet, the decision-making process behind migrants choosing their destination remains elusive. Existing models primarily rely on population size and travel distance to explain the spatial patterns of migration flows, overlooking significant population heterogeneities. Paradoxically, migrants often travel long distances and to smaller destinations if their diaspora is present in those locations. To address this gap, we propose the diaspora model of migration, incorporating intensity (the number of people moving to a country), and assortativity (the destination within the country). Our model considers only the existing diaspora sizes in the destination country, influencing the probability of migrants selecting a specific residence. Despite its simplicity, our model accurately reproduces the observed stable flow and distribution of migration in Austria (postal code level) and US metropolitan areas, yielding precise estimates of migrant inflow at various geographic scales. Given the increase in international migrations, this study enlightens our understanding of migration flow heterogeneities, helping design more inclusive, integrated cities.

Read the full article at: academic.oup.com

Iacopo Iacopini, Jennifer R. Foote, Nina H. Fefferman, Elizabeth P. Derryberry and Matthew J. Silk

Phil Trans Roy Soc B

08 July 2024 Volume 379Issue 1905

Animal communication is frequently studied with conventional network representations that link pairs of individuals who interact, for example, through vocalization. However, acoustic signals often have multiple simultaneous receivers, or receivers integrate information from multiple signallers, meaning these interactions are not dyadic. Additionally, non-dyadic social structures often shape an individual’s behavioural response to vocal communication. Recently, major advances have been made in the study of these non-dyadic, higher-order networks (e.g. hypergraphs and simplicial complexes). Here, we show how these approaches can provide new insights into vocal communication through three case studies that illustrate how higher-order network models can: (i) alter predictions made about the outcome of vocally coordinated group departures; (ii) generate different patterns of song synchronization from models that only include dyadic interactions; and (iii) inform models of cultural evolution of vocal communication. Together, our examples highlight the potential power of higher-order networks to study animal vocal communication. We then build on our case studies to identify key challenges in applying higher-order network approaches in this context and outline important research questions that these techniques could help answer.

Read the full article at: royalsocietypublishing.org

Guy Amichay, Liang Li, Máté Nagy & Iain D. Couzin 

Nature Communications volume 15, Article number: 4356 (2024)

Coordinated motion in animal groups has predominantly been studied with a focus on spatial interactions, such as how individuals position and orient themselves relative to one another. Temporal aspects have, by contrast, received much less attention. Here, by studying pairwise interactions in juvenile zebrafish (Danio rerio)—including using immersive volumetric virtual reality (VR) with which we can directly test models of social interactions in situ—we reveal that there exists a rhythmic out-of-phase (i.e., an alternating) temporal coordination dynamic. We find that reciprocal (bi-directional) feedback is both necessary and sufficient to explain this emergent coupling. Beyond a mechanistic understanding, we find, both from VR experiments and analysis of freely swimming pairs, that temporal coordination considerably improves spatial responsiveness, such as to changes in the direction of motion of a partner. Our findings highlight the synergistic role of spatial and temporal coupling in facilitating effective communication between individuals on the move.

Read the full article at: www.nature.com