Over the past decades, the cultural evolution field has generated an important body of knowledge using experimental, historical, and computational methods. While these approaches have generated testable hypotheses, many phenomena are too complex for agent-based models. Here we propose that an employ of Large Language Models (LLMs) can be a novel way to represent human behavior. We simulate cultural evolution in populations of LLMs by variables such as network structure, personality, and social information. The software for conducting these simulations is open-source and features a user-interface to build bridges between the fields of cultural evolution and generative artificial intelligence.

The Flowers project-team, at the University of Bordeaux and at Ensta ParisTech, studies versions of hoistic individual development. These models can help us better understand how children learn, as well as to build machines that gain knowledge as children do, aka developmental artificial intelligence, with applications in educational technologies, automated discovery, robotics and human-computer interaction.

Yang, Zhaohui and Kshitji Jerath. Multi-scale Traffic Flow Modeling: A Renormalization Group Approach. arXiv: 2403.13779. UMass Lowell engineers achieve a unique advancement in the mathematical study of human mobilities from the viewpoint of a significant physical phenomena, as the title notes. At once the work identifies this double dimension and serves to trace and connect our Earthuman travels with universal principles.

Traffic flow modeling is typically performed at one of three (microscopic, mesoscopic, or macroscopic) scales. Recent works to merge models have had some success, but a need still exists for a single framework that can model traffic flow across spatiotemporal phases. Here we utilize a renormalization group (RG) theoretic approach, building upon our prior research on statistical mechanics-inspired traffic flow studies. We measure the coarse-grained traffic flow simulation using a pixel-based image metric and find good correlation in each case. (Excerpt)

In theoretical physics, the term renormalization group refers to the systematic investigation of changes of a physical system as viewed at different scales. The renormalization group is intimately related to scale invariance and conformal invariance, symmetries in which a system appears the same at all scales (so-called self-similarity). (Wikipedia)

My overarching research goal is to advance the understanding of complex dynamics observed in large-scale self-organizing systems, and to design bottom-up control algorithms that guide such systems to desired states via minimal intervention. (K. Jerath)

Ribiero, Fabiano and Vinicius Netto. Urban Scaling Laws. arXiv:2404.02642. In this chapter for a Compendium of Urban Complexity, a Universidade Federal de Lavras (UFLA), Brazilphysicist and a Research Centre for Territory, Transports and Environment (CITTA), University of Porto, Portugal architect chronicle the structural presence of self-similarities across large and smaller human habitations.

Understanding how size influences the internal characteristics of a system is a crucial concern. Concepts like scale invariance, universalities, and fractals find application in biology, physics, and particularly urbanism. Relative size impacts how cities form and function economically and socially. For example, what are the pros and cons of larger cities? Do they offer more opportunities and higher incomes than smaller ones? To address such issues, we utilize theoretical tools from scaling theory to quantify how a system's behavior changes across macro to micro scales. Drawing parallels with biology and spatial economics, this chapter explores recent discoveries, ongoing progress, and new questions regarding urban scaling.

Yadav, Pawanesh, et al. Explaining Indian Stock Market through Geometry of Scale free Networks. arXiv:2404.04710. We note this work by Shiv Nadar University, Uttar Pradesh, India mathematicians as an example of how every natural and social domain can be traced to and expressed in complex dynamical system theories. In regard then, these findings once again serve define a double manifest and prescriptive biological reality.

This paper presents an analysis of the Indian stock market using a method based on embedding the network in a hyperbolic space using Machine learning techniques. We claim novelty on four counts. First, the hyperbolic clusters resemble the topological network communities better the Euclidean clusters. Second, we clearly distinguish between periods of market stability and volatility through a statistical analysis of distance and shortest path corresponding to the embedded network. Third, we use the modularity of the embedded network so market changes can be spotted early. Lastly, our approach segregate market sectors thereby underscoring its natural clustering ability. (Abstract)

Earth Earns: An Open Participatory Earthropocene to Astropocene CoCreative Future

Jaderberg, Ben, et al. Potential of quantum scientific machine learning applied to weather modelling.. arXiv:2404.08737. Eight theorists at PASQAL, Massy, France, BASF Digital, Ludwigshafen, Germany and BASF Research Park, North Carolina report a first ever mid 2020s combine of quantum phase abilities by way of AI neural net methods to reach a worldwise achievement in climate analyses and forecasts.

In this work we explore how quantum machine learning can advance the science of weather modelling. Using quantum circuits as a device, we consider supervised learning from weather data and physics-informed solving of atmospheric dynamics. In the first case, a quantum model can be trained on global stream function dynamics. We then introduce the barotropic vorticity equation (BVE) as our model of the atmosphere. Using the quantum circuits algorithm, we solve the BVE under boundary conditions and use the trained model to predict unseen future weather states. Whilst challenges remain, our results mark an advancement in terms of the complexity of PDEs solved with quantum scientific machine learning. (Excerpt)

PASQAL is a leading Quantum Computing company in France that builds quantum processors from ordered neutral atoms in 2D and 3D arrays to bring a practical quantum advantage. PASQAL was founded in 2019, out of the Institut d'Optique, by Georges-Olivier Reymond, Christophe Jurczak, Alain Aspect, Nobel Physics, 2022, Antoine Browaeys, and Thierry Lahaye.

Gianfrate, Antonio, et al. Reconfigurable quantum fluid molecules of bound states in the continuum. Nature Physics. 20/1, 2024. We enter this work by thirteen nanoscientists mainly at CNR Nanotechnology, Italy and Princeton University as another instance of mid 2020s Earthuman abilities to learn all about and delve into any depth of as an evolitionary project to begin an new intentional quantum phase cocreation.

Topological bound states are confined wave-mechanical objects that offer advantageous ways to enhance light–matter interactions in photonic devices. Here we show that polariton condensation into a negative-mass bound state in the continuum exhibits interactive confinement to attain optically reprogrammable molecular arrays of quantum fluids of light. We demonstrate the scalability of our technique by extended mono- and diatomic chains of bound-state-in-the-continuum polariton fluids. (Excerpt)

Mitra, Anupam, et al.. Macrostates vs. Microstates in the Classical Simulation of Critical Phenomena in Quench Dynamics of 1D Ising Models. arXiv:2310.08567.. This entry by Center for Quantum Information and Control, University of New Mexico physicists including Ivan Deutsch is posted as an example among many to show how readily human intellects can delve into these fundamental depths and then to take over and commence anew a second intentional, informed material cocreation. See also Ultracold field-linked tetratomic molecules by Chen, Xing-Yan Chen, et al in Nature (January 31, 2024) for a similar instance.

We study the tractability of classically simulating critical phenomena in the quench dynamics of one-dimensional transverse field Ising models (TFIMs) using highly truncated matrix product states (MPS). We focus on two paradigmatic examples: a dynamical quantum phase transition (DQPT) that occurs in nonintegrable long-range TFIMs, and the infinite-time correlation length of the integrable nearest-neighbor TFIM when quenched to the critical point. For the DQPT, we show that the order parameters can be efficiently simulated with surprisingly heavy truncation of the MPS bond dimension. This can be used to reliably extract critical properties of the phase transition, including critical exponents, even when the full many-body state is not simulated with high fidelity.

Terasa, Ivo, et al. Pathways towards truly brain-like computing primitives. MaterialsToday. October, 2023. In this Springer journal, twelve Institute of Materials Science, Kiel University, Germany researchers describe a novel application of AI neural large learning and an avail of animate complex system principles such as Distributed Plasticity: Operation near Criticality, Self-ordered Arrangement, Hierarchy, Modularity, Robustness, and Oscillatory Ensembles. In this regard, the team broaches an innovative synthesis of these methods and features by which to open a 2020s frontier of intentional cocreativity going forward.

Taking inspiration from biological and neural information processing, deep learning and artificial intelligence have made solutions to complex problems more feasible. To explore the capabilities of brain-like hardware computing, a platform with dynamic reconfigurable connections is mandatory. This work addresses this biological motivation and classifies them with respect to several fundamental principles of brain-like computing. The approaches range from interconnected nanogranular networks with dynamically reconfigurable connections and guided redox-wiring to the mimicking of neural action potentials by relaxation-type oscillators that are used as input stimuli. (Abstract excerpt)

The nervous systems of humans, mammals, and even simple living species like invertebrates are well adapted to changing environments. The remarkable interactions with their surroundings are a result of millions of years of evolution. Biological systems can perform cognitive tasks, such as pattern recognition, with low power consumption. Fundamental building blocks leading to such core abilities exploit neurons as central processing units, which are interconnected by synapses to form a complex dynamical three-dimensional network. It is therefore no surprise that attempts have been made to develop artificial information processing systems, to reach the performance and power efficiency of nervous systems. Machine Learning, Neuromorphic Engineering and in materia Computing can be identified as three main development avenues (1, 2)

Previous   1 | 2 | 3 | 4 | 5 | 6  Next