News + Events

Economic inequality is one of our primary global challenges and is a key research topic for archaeology — Why do some societies become deeply unequal while others remain more balanced? What clues about our economic past are hidden in the ruins of ancient homes? A new Special Feature in PNAS, edited by SFI External Professors Tim Kohler (Washington State University), Amy Bogaard (University of Oxford), and facilitated by Scott Ortman (University of Colorado Boulder), highlights papers by archaeologists, anthropologists, historians, and economists exploring these questions using a global database of over 53,000 residential buildings from about 4,000 archaeological settlements.

A recent study by SFI Postdoctoral Fellow Andrew Stier and Professor Brandon Ogbunu in NPJ Complexity shows that many U.S. cities diverge from common narratives about political leanings and implicit bias.

In a recent analysis, SFI Complexity Postdoctoral Fellow Yuanzhao Zhang and collaborator William Gilpin reported that one foundation model called Chronos could generate predictions of chaotic dynamical systems at least as accurately as models trained on relevant data. The team presented their work at the Thirteenth International Conference on Learning Representations, saying the paper represents the first test of zero-shot learning in forecasting chaotic systems.

Five millennia ago, wealth inequality — which had stayed roughly constant for thousands of years — exploded. It has stayed constant, albeit much higher, ever since. One factor was the ox-drawn plow, according to a recent paper in the Journal of Economic Literature. 

Collective action can be highly effective within a community, but can be harder to achieve across communities. People who have friends across communities could help tackle this challenge, finds a recent study in Conservation Letters, in which SFI External Professor Monique Borgerhoff Mulder and colleagues surveyed residents across 28 Tanzanian fishing villages about participation in beach management activities.

Humans learn by breaking through and plateauing, persisting and resting, and, occasionally, experiencing the blissful flow state. Mastering a skill can take decades, but the learning process unfolds across multiple timescales, from mere moments to days. A new study presents a unified theoretical model, capturing the different timescales of learning.

Scientists usually use a hypergraph model to predict dynamic behaviors. But the opposite problem is interesting, too. What if researchers can observe the dynamics but don’t have access to a reliable model? Yuanzhao Zhang, an SFI Complexity Postdoctoral Fellow, has an answer. In a paper published in Nature Communications, Zhang and his collaborators describe a novel algorithm that can infer the structure of a hypergraph using only the observed dynamics.

In many careers, a person must learn foundational skills before advancing more deeply into their profession. A recent paper in Nature Human Behavior mapped the dependency relationships between workplace skills, identifying a nested structure in many professions, where advanced skills depend on prior mastery of broader skills. This nestedness, they found, has significant implications for wage inequality and career mobility in increasingly complex labor markets.

When comparing two objects, people either rely on internal memories of these objects or run their hands and eyes over them to directly perceive their similarity. The latter approach, a shortcut that offloads cognition to the active perceptual operations like eye or hand movements, requires a lower memory burden. In a study published in the journal Attention, Perception, & Psychophysics, SFI Complexity Postdoctoral Fellow Marina Dubova and SFI Research Fellow Arseny Moskvichev demonstrate that it is also more effective.

A new paper in PNAS shows that the idea of “taking turns” could help resolve the 1960s paradox of the plankton — and better predict how climate change will remake our oceans.

Using a variety of analogy puzzles, SFI researchers have shown that the reasoning abilities of OpenAI’s GPT-4 model fall short when faced with small changes. Recognizing the limitations of these tools is critical to knowing when and to what extent they can be trusted.

Cultural traits — the information, beliefs, behaviors, customs, and practices that shape the character of a population — are influenced by conformity, the tendency to align with others, or anti-conformity, the choice to deliberately diverge. A new way to model this dynamic interplay could ultimately help explain societal phenomena like political polarization, cultural trends, and the spread of misinformation. 

Knowing only the building blocks of our own biosphere, can we predict how life may exist on other planets? What factors will rein in the Frankensteinian life forms we hope to build in laboratories here on Earth? A paper in Interface Focus co-authored by several SFI researchers takes these questions out of the realm of science fiction and into scientific laws.

Multi-scale complex systems are ubiquitous and also notoriously difficult to model. In disturbed systems, conventional bottom-up or top-down approaches can’t capture the interactions between the small-scale behaviors and the system-level properties. SFI External Professor John Harte and his collaborators have worked to resolve this challenge by building a hybrid method that links bottom-up behaviors and top-down causation in a single theory. 

A recent study by SFI Complexity Postdoctoral Fellow Katrin Schmelz and coauthors explores various scenarios when it may be beneficial to stand out and when it might be better to blend in. 

In a new PNAS Perspective, a team of authors argues that stochastic thermodynamics provides the mathematical tools needed to investigate the energy usage of all computational systems.

A new study in Theory and Society shows that the printing of witch-hunting manuals, particularly the Malleus Maleficarum in 1487, played a crucial role in spreading persecution across Europe. The study also highlights how trials in one city influenced others. This social influence — observing what neighbors were doing — played a key role in whether a city would adopt witch trials.

In the 1990s, a new computational model called AlChemy promised to explain where we came from — why specific basic chemical compounds combine to form networks as complex as life. A new paper revives the AlChemy model, reporting that it predicts stable complex organizations (like life) will emerge from basic parts more frequently than previously believed. The article appears in a special issue of Chaos, co-edited by SFI External Professor Elizabeth Bradley and celebrating Science Board Fellow David Campbell.

In a new paper in Science Advances, SFI Complexity Postdoctoral Fellow Yuanzhaho Zhang reports on how higher-order interactions can reshape a system at large — even global — scales.

The beliefs we hold develop from a complex dance between our internal and external lives. A recent study published in Psychological Review uses well-known formalisms in statistical physics to model multiple aspects of belief-network dynamics. This multidimensional approach to modeling belief dynamics could offer new tools for tackling various real-world problems such as polarization or the spread of disinformation.