Dynamical Phase Transitions in Non-equilibrium Networks

Jiazhen Liu, Nathaniel M. Aden, Debasish Sarker, Chaoming Song · 2024

Read Dynamical Phase Transitions in Non-equilibrium Networks by Jiazhen Liu, Nathaniel M. Aden, Debasish Sarker, Chaoming Song (2024) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2412-06704.txt.

Introduction

Liu and co-authors build a lone form for phase shifts in off-rest webs in the work. Phase shifts at set times mark sharp shifts in web acts. They give a lens on off-rest acts past rest-frame rules in past work. Such shifts got wide study in quant nets in past years. In prime nets they stayed near bare till this work came. Late tests on hard nets show sharp shifts like those in quant nets.

Those nets span social webs and cash shops in test runs. Their form ties bent edge ties to such shifts in webs. Web grade swells at a set time in that form's runs. A wide bell law marks that swell in the runs there. Key web marks, like grade shares, scale at the edge in runs. Ends ground a base for wild off-rest peaks in hard nets.

Catalog record

CREATORJiazhen Liu, Nathaniel M. Aden, Debasish Sarker, Chaoming Song
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE24 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-06704

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Dynamical Phase Transitions in Non-equilibrium Networks Jiazhen Liu, Nathaniel M. Aden, Debasish Sarker, Chaoming Song Abstract Dynamical phase transitions (DPTs) characterize critical changes in system behavior occurring at finite times, providing a lens to study nonequilibrium phenomena beyond conventional equilibrium physics. While extensively studied in quantum systems, DPTs have remained largely unexplored in classical settings. Recent experiments on complex systems, from social networks to financial markets, have revealed abrupt dynamical changes analogous to quantum DPTs, motivating the search for a theoretical understanding. Here, we present a minimal model for nonequilibrium networks, demonstrating that nonlinear interactions among network edges naturally give rise to DPTs. Specifically, we show that network degree diverges at a finite critical time, following a universal hyperbolic scaling, consistent with empirical observations. Our analytical results predict that key network properties, including degree distributions and clustering coefficients, exhibit critical scaling as criticality approaches. These findings establish a theoretical foundation for understanding emergent nonequilibrium criticality across diverse complex systems. Dynamical phase transitions (DPTs) are critical phenomena that unveil critical behavior emerging over time and represent a new frontier of nonequilibrium physics [ 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 ] . Distinct from conventional phase transitions, typically driven by external parameters such as temperature or pressure, DPTs occur at a critical time t_{c} , where physical observables display nonanalytic behavior. While these transitions have been extensively studied in quantum systems [ 7 , 8 , 9 ] , their role in classical contexts remains largely underexplored. Intriguingly, recent experiments on macroscopic complex systems, including ecological collapse [ 10 , 11 ] , financial crises [ 12 , 13 ] , and power grid failures [ 14 ] , report abrupt dynamical changes at finite times. These events, though not formally classified as DPTs, share key characteristics such as abrupt, collective changes. The striking similarity between these phenomena and quantum DPTs suggests that critical temporal dynamics may be a widespread feature of complex systems. This raises intriguing questions about their broader relevance and underlying universality. More recently, studies on social networks have presented compelling evidence for genuine DPTs in classical settings [ 15 , 16 , 17 , 18 ] . These experiments reveal some physical quantities z(t) diverges at a critical time t_{c} according to a universal hyperbolic scaling: z(t)\sim(t_{c}-t)^{-1}. (1) Unlike quantum DPTs, where nonanalyticity emerges without divergence, this classical hyperbolic scaling reveals a distinct signature of temporal criticality. Moreover, this hyperbolic divergence appears to be ubiquitous across distinct domains. For instance, hyperbolic discounting observed in economics [ 19 , 20 , 21 , 22 ] , which rationalizes procrastination and deadlines, also adheres to this scaling, as evidenced by surges in academic submissions near deadlines [ 23 ] . Despite its prevalence, the theoretical underpinnings of this universality remain poorly understood. Existing theoretical frameworks for nonequilibrium classical phase transitions generally fit into two broad categories. The first addresses different phases emerging under time-varied control parameters, much like conventional equilibrium transitions [ 24 , 25 , 26 , 8 , 4 , 27 , 28 , 29 , 30 , 31 ] . Some studies even allow these control parameters to vary over time, promoting the system through distinct dynamical regimes [ 32 , 17 ] . The second encompasses models of self-organized criticality [ 33 , 34 , 35 , 36 , 37 , 38 , 39 ] , where critical points occur in the asymptotic limit t\to\infty rather than at a finite t_{c} . Neither category fully captures the finite-time singularities characteristic of DPTs, indicating a need for new theoretical approaches. In this study, we introduce a simple theoretical framework that predicts the emergence of DPTs in nonequilibrium networks. Our central insight is that nonlinear interactions play a pivotal role. Drawing an analogy to many-body physics, a random graph without interactions behaves like a system of noninteracting particles with no DPTs. Once nonlinear interactions are introduced, however, the system’s collective dynamics are fundamentally altered. To demonstrate this, we consider a simple model involving quadratic interactions, motivated by triangle closure, a ubiquitous mechanism in social networks. Within this model, quadratic interactions between edges naturally yield DPTs. Remarkably, the model can be solved analytically, reproducing the universal scaling behavior expressed in Eq. ( 1 ). Furthermore, we derive critical network characteristics, such as the degree distribution and clustering coefficient, exhibit critical scaling behavior as t_{c} is approached. These results establish a theoretical basis for understanding nonequilibrium criticality in classical systems, shedding light on parallels with quantum DPTs. Figure 1: Illustration of network evolution over time for \beta=0.5 , \alpha=1 , and \gamma=1 with N=300 : ( a ) Before the critical time t_{c} , the network exhibits a sparse structure. ( b ) At t\to t_{c} , the network transitions to a scale-free topology. ( c ) After t_{c} , the network becomes dense, resembling a nearly complete graph. Theorical framework We consider a directed network comprising N nodes undergoing random edge addition and removal. Specifically, a node i connects to a node j to form a new edge with a probability rate \alpha/N , while edges are removed at a rate \gamma . The parameters \alpha and \gamma characterize the rates of edge formation and removal, respectively, and the factor 1/N ensures the sparsity of the network. The evolution of the expected average degree, z(t) , follows the linear differential equation: {\frac{d{z}}{d{t}}}=\alpha-\gamma z, with the stationary solution given by z(t\to\infty)=\alpha/\gamma . In this non-interacting scenario, the network remains a random graph, exhibiting no DPTs. To induce a DPT, it is necessary to introduce interactions. A natural approach is to modify the edge addition rate to \alpha/N+U_{int} , where U_{int}(A) represents an interaction term dependent on the Boolean adjacency matrix A . While the edge removal rate could also be modified similarly, we hold this parameter constant for simplicity. The interaction term can be expanded as U_{int}(A)=\beta_{0}A+\beta A^{2}+O(A^{3}) . To isolate the nonlinear effects, we neglect the linear term and retain only the leading nonlinear contribution, yielding U_{int}(A)\approx\beta A^{2} , where \beta is the coupling constant. These interactions are naturally interpreted as a triadic closure mechanism: two nodes, i and j , sharing a common neighbor are more likely to form a direct connection. This occurs with an additional rate proportional to \sum_{k}A_{ik}A_{kj}=(A^{2})_{ij} . While triadic closure is typically associated with social networks, our theory treats it as the leading nonlinear interaction, extending its applicability beyond social contexts. Figure 2: Theoretical predictions versus numerical simulations for ( a ) the average degree z(t) and ( b ) the order parameter q(t) , with parameters \alpha=1 , \gamma=1 , \beta=1/4 (green), and \beta=1/2 (red). Solid curves represent theoretical predictions, while scatter points correspond to simulations for N=2,000 . For \beta=1/4 (green), the average degree z(t) converges, while the order parameter q(t) vanishes, indicating an equilibrium phase. For \beta=1/2 (red), z(t) exhibits hyperbolic scaling and diverges at a finite critical time t_{c}\approx 5.9 . Correspondingly, q(t) shows a sharp jump at t_{c} , signifying a first-oder DPT from a sparse to a dense network. In the weak-interaction regime ( \beta small), the network behaves as a random graph with no DPTs.

Read the complete text