SigmaPhi 2023

Workshop organized by:
- D. Hristopulos and D. Valenti
- S. Blesic
- P. Ditlevsen, M. Ghil, N. Boers and M. Rypdal

Section I: Natural Systems, Complexity and Environmental Modeling: The Triangle of Statistical Physics, Statistics, and Machine Learning.

D. Hristopulos and D. Valenti

The aim of this workshop is to bring together contributions on theoretical, experimental, and computational approaches for studying the complexity in natural systems by exploiting tools inspired by statistical physics, spatio-temporal statistics, and machine learning. Statistical approaches and machine learning methods ---which have strong links with statistical physics--- are used to analyze and extract information from complex patterns in environmental data. This workshop aims to highlight such connections and to present novel ideas and methods motivated by statistical physics that can lead to new environmental applications and insights. Both stochastic methods and approaches based on the theory of dynamical systems are welcome. A non-exclusive list of topics of interest includes novel computational and theoretical tools for the analysis of large spatio-temporal data sets, modeling of natural systems as intrinsically nonlinear open systems, methods that address multiple-scale interactions, approaches for the reconstruction and simulation of natural or engineered porous media with non-Gaussian statistics, applications of stochastic differential equations to environmental processes, higher-order upscaling methods, applications of complex network theory, estimation of long-range correlations in environmental systems. Physical phenomena of interest include (but are not limited to): the flow and transport of pollutants in the atmosphere, oceans and subsurface, natural hazards (earthquakes, fires, avalanches, and landslides), heat waves and precipitation.

Section I: Understanding climate, contributing to overall adaptation efforts

S. Blesic

The most pressing issues facing current interdisciplinary efforts that deal with the complexity of climate change are in advancements in understanding and explaining the physical basis of climate dynamics realized, in parallel, with utilization of that knowledge to effectively contribute to new lines of research that will develop innovative applications that drive particularly adaption efforts. This workshop will be organized to showcase current research and research potential of both paths. Statistical physics community has already done a lot of important work in understanding climate variability. Therefore, the workshop will present statistical physics approaches to understanding of physical aspects of climate phenomena. The workshop will be extended with work that investigates various physical and non-physical phenomena in climatic context, thus contributing to the overall adaptation efforts. These kinds of investigations have an additional appeal, particularly to statistical physics community that standardly works in interdisciplinary areas, for they usually bring together researchers, practitioners and communities with discrete expertise, in order to better understand climate change and its impacts. Focuses of these researches are different but areas of interest are vast and can include any critical issue that climate change threatens to seriously exacerbate. What statistical physics can offer in this context is data- or model-led understandings that are of wider value to the scientific community and applicable local-scale insights.

Section II:Critical transitions and climate change

P. Ditlevsen, M. Ghil, N. Boers and M. Rypdal

Several components of the climate system have been identified as possessing a potential risk for undergoing abrupt transitions; such components have been called Tipping Elements (TEs). The interaction between different components in the complex Earth system could lead to a cascade of tipping events, with the probability of critical transitions within one TE depending on the evolving state of one or more other TEs. Understanding this kind of cascading behavior and the phenomena underpinning the tipping events involved requires use of statistical physics tools to understand critical transitions in complex systems. Such tools are provided by the theory of fast-slow systems, dynamical and stochastic systems theory, nonlinear time series analysis, and multiple time-scale dynamics. The applications include investigations of paleoclimatic records and present dayâ€™s observations, as well as the behavior of TEs in Earth System Models, where computer simulations must be carefully designed to explore the possible transitions.

In this symposium, we invite contributions that further develop and apply methods from statistical physics. Particular emphasis will be placed on the study of climate response to increased greenhouse gas concentrations, climate tipping points, time-dependent forcing and associated pullback attractors in climate evolution, as well as extreme and rare events in observations and models and the uses of statistical mechanics across the hierarchy of Earth System Models.

Workshop organized by: B. Tadic and N. Gupte

The influence of network structure on dynamics in many complex systems has been demonstrated in numerous studies with a detailed analysis of empirical data and theoretical approaches. Recent studies of networks representing various complex systems from the brain to large-scale social dynamics have revealed their higher-order architecture, which can be described as aggregates of simplexes (triangles, tetrahedrons, and higher cliques). Beyond the standard graph theory, these hidden structures are quantified by algebraic topology methods. This Special Session will bring together presentations exploring different complex systems concerning a) the structure of simplicial complexes, including the properties of the underlying topological graph (network); b) dynamic processes in a particular or co-evolving network structure; c) information topology and graph representations of time series. We expect research results based on empirical data analysis and theoretical and numerical modeling. The aim is to cover recent developments in these areas of research and to stimulate discussion towards an in-depth understanding of the role of higher-order connectivity in the emergence of functional properties of complex systems.

Workshop organized by: P. Argyrakis and T. Di Matteo

Workshop webpage

During recent decades, the financial market landscape has changed dramatically with the deregulation of markets and the growing complexity of products. The ever-increasing speed and decreasing costs of computational power and networks have led to the emergence of huge datasets. Data science is the discipline that deals with collecting, praparing, managing, analysing, interpreting and visualising large and complex datasets. The availability of these datasets permits the development of data driven models that are developed on the empirical observations of economic and financial systems. Despite data driven models are popular approaches, there has been little interaction between these techniques and econophysics modelling.

Econophysicists mainly apply methods from statistical physics, the physics of complex systems and science of networks to macro/micro-economic modelling, financial market analysis and social problems and they mainly ground their studies on empirical evidences.

The present workshop aims to bridge that gap between data driven and econophysics modelling by inviting submissions that apply and ideally combine these techniques to economics and finance systems in real-world applications.

Workshop organized by: F. Caruso

This workshop will cover the fascinating very interdisciplinary field of machine learning and its more recent generalization to quantum information. Recurrent networks, reinforcement learning, Boltzmann machines, deep learning, and neural networks will be discussed in various scientific fields and from both foundational and technological perspectives.

On one side, machine learning does very successfully solve very hard problems as detecting anomalous events in live streams of sensor data, self-driving cars, speech recognition, natural language, vision, playing games, trading strategies, etc.

On the other side, quantum information science also teems with excitement. By manipulating single particles at a subatomic level, one is able to perform Fourier transformation exponentially faster or search in a database quadratically faster than the classical limit. A future quantum computer will break the security of all the currently encrypted messages, and quantum cryptography already shows instead of an unbreakable tool to faithfully transmit information at a commercial level.

But, what can quantum computing contribute to machine learning? We naturally expect a speedup from quantum methods, but what kind of speedup?

We will bring together experts from quantum computing and machine learning to discuss the latest progress in the rapidly growing field of quantum machine learning.

ΣΦ Conference

W4. Climate and Environments

W3. Complex Networks: Hidden Geometry and Dynamics

W2. Data Science and Econophysics

W1. Quantum Physics and Machine Learning

Workshops

W1. Quantum Physics and Machine Learning

W2. Data Science and Econophysics

W3. Complex Networks: Hidden Geometry and Dynamics

W4. Climate and Environments

W5. Statistical Physics of Biophysical Systems

W6. Statistical Physics of Glasses

W7. Fluctuation-Dissipation Theorem

W8. Phase Transition and Topological Phenomena

W9. Non-Extensive Statistical Mechanics and Kappa Distributions