SCTPLS 35: A few more weeks!

[Guastello, Stephen]

A few more weeks until the 35th Annual International Conference of the SCTPLS!!

July 30th - August 1st, 2025, at the University of Colorado, Colorado Springs, CO

If you haven't registered already, today (7/10/2025) is the last day for the early bird rate!

https://www.societyforchaostheory.org/conf/2025/cfp/

You won't want to miss Dr. Aaron Clauset and Dr. Nicholas Stergiou as our keynote speakers at this years conference

 

Dr Aaron Clauset will present:

Nearly-Optimal Prediction of Missing Links in Networks

Networks are ubiquitous in real-world data applications, e.g., in social network analysis or biological modeling, but networks are also nearly always incompletely observed. Current algorithms for predicting missing links in the hard case of a network without node attributes exhibit wide variations in their accuracy, and we lack a general understanding of how algorithm performance depends on the input network's characteristics. In this talk, I'll describe a powerful meta-learning solution to this problem that makes nearly-optimal predictions by learning to combine or 'stack' many individual link prediction methods. Using synthetic data for which we can analytically calculate the optimal performance and a large corpus of 550 structurally diverse networks from social, biological, technological, information, economic, and transportation domains, we systematically evaluate more than 200 link prediction methods individually and in combined stacked models. Across most settings, we show that model stacking nearly always performs best and produces nearly-optimal performance on synthetic networks. Furthermore, compared to state-of-the-art graph neural network techniques, we find that model stacking is typically more computationally efficient and equally accurate on multiple measures of performance. Applied to real networks, we find that the difficulty of predicting missing links varies considerably across domains: it is easiest in social networks and hardest in economic and biological networks, but performance depends strongly on network characteristics like the degree distribution, triangle density, and degree variation. I'll close with some commentary on future work on link prediction problem.

Dr. Nicholas Stergiou will present:

Variability in Movement

A large body of research demonstrates the existence of an optimal level of variability which enables us to interact adaptively and safely to a continuously changing environment, where often our movements must be adjusted in a matter of milliseconds. Decrease, or loss of this optimal level due to neurodegenerative and physiological disorders makes the system more rigid and less adaptable to different perturbations. Increase makes the system noisier and more unstable. Stable behavior is a rich behavioral state with high complexity, where complexity is defined as highly variable fluctuations in physiological processes resembling mathematical chaos and fractals thus being more nature based. In this keynote, I present updates of this field of research regarding the innovative “next step” that goes beyond the many descriptive studies that characterize levels of variability in various populations. This research aims to eventually devise novel interventions and technologies that will harness the existing knowledge on variability and create new possibilities for those in need to improve performance and/or restore their decreased physical abilities.

Reminders:

We are looking forward to seeing you all there!

https://www.societyforchaostheory.org/conf/2025/

 

Cordially yours, the Conference Committee,

Adam Kiefer, University of North Carolina at Chapel Hill, (SCTPLS President-Elect and Conference Chair)

Barney Ricca, University of Colorado, Colorado Springs (SCTPLS President)

Charles Benight, University of Colorado, Colorado Springs

Cortney Armitano-Lago, University of North Carolina at Chapel Hill

Stephen Guastello, Marquette University, Milwaukee, WI