Interplay between high-order structures and contagion dynamics

Recently, the learning by confusion (LbC) approach has been proposed as a machine learning tool to determine the critical temperature T_{c} of phase transitions without any prior knowledge of its even approximate value. The method has been proven effective, but it has been used only for continuous phase transitions, where the confusion results only from deliberate incorrect labeling of the data. However, in the case of a discontinuous phase transition, additional confusion can result from the coexistence of different phases. To verify whether the confusion scheme can also be used for discontinuous phase transitions, we apply the LbC method to three microscopic models, the Blume-Capel, the q-state Potts, and the Falicov-Kimball models, which undergo continuous or discontinuous phase transitions depending on model parameters. With the help of a simple model, we predict that the phase coexistence present in discontinuous phase transitions can indeed make the neural network more confused and thus decrease its performance. However, numerical calculations performed for the models mentioned above indicate that other aspects of this kind of phase transition are more important and can render the LbC method even less effective. Nevertheless, we demonstrate that in some cases the same aspects allow us to use the LbC method to identify the order of a phase transition.

Synergistic contagion in a networked system occurs in various forms in nature and human society. While the influence of network's structural heterogeneity on synergistic contagion has been well studied, the impact of individual-based heterogeneity on synergistic contagion remains unclear. In this work, we introduce individual-based heterogeneity with a power-law form into the synergistic susceptible-infected-susceptible model by assuming the synergistic strength as a function of individuals' degree and investigate this synergistic contagion process on complex networks. By employing the heterogeneous mean-field (HMF) approximation, we analytically show that the heterogeneous synergy significantly changes the critical threshold of synergistic strength σc that is required for the occurrence of discontinuous phase transitions of contagion processes. Comparing to the synergy without individual-based heterogeneity, the value of σc decreases with degree-enhanced synergy and increases with degree-suppressed synergy, which agrees well with Monte Carlo prediction. Next, we compare our heterogeneous synergistic contagion model with the simplicial contagion model [Iacopini et al., Nat. Commun. 10, 2485 (2019)], in which high-order interactions are introduced to describe complex contagion. Similarity of these two models are shown both analytically and numerically, confirming the ability of our model to statistically describe the simplest high-order interaction within HMF approximation.

The phenomenon of load redistribution in complex networks has garnered extensive attention due to its profound impact and widespread occurrence. In recent years, higher-order structures have offered new insights into understanding the structures and dynamic processes of complex networks. However, the influence of these higher-order structures on the dynamics of load redistribution, cascade failures, and recovery processes remains to be fully explored. In this study, we propose the load redistribution model with higher-order structures and recovery strategies of cascade failure based on functional upgrading and reconstruction mechanisms. In the cascading failure process with load redistribution and higher-order recovery strategies, we find that higher-order structures can induce a discontinuous phase transition at the low proportion of load redistribution, and the dynamic process displays a dual character of being robust yet fragile. These findings have been examined in both real and classical modeled networks. Interestingly, the largest connected component exhibits three distinct modes as the attack ratio increases at high densities of higher-order structures and recovery mechanisms.

Lateral predictive coding is a recurrent neural network that creates energy-efficient internal representations by exploiting statistical regularity in sensory inputs. Here, we analytically investigate the trade-off between information robustness and energy in a linear model of lateral predictive coding and numerically minimize a free energy quantity. We observed several phase transitions in the synaptic weight matrix, particularly a continuous transition that breaks reciprocity and permutation symmetry and builds cyclic dominance and a discontinuous transition with the associated sudden emergence of tight balance between excitatory and inhibitory interactions. The optimal network follows an ideal gas law over an extended temperature range and saturates the efficiency upper bound of energy use. These results provide theoretical insights into the emergence and evolution of complex internal models in predictive processing systems.

An infinite array of globally coupled overdamped constituents moving in a double-well potential with nth order saturation term under the influence of additive Gaussian white noise is investigated. The system exhibits a continuous phase transition from a symmetric phase to a symmetry-broken phase. The qualitative behavior is independent on n. The critical point is calculated for strong and for weak noise; these limits are also bounds for the critical point. Introducing an additional nonlinearity, such that the potential can have up to three minima, leads to richer behavior. There the parameter space divides into three regions: a region with a symmetric phase, a region with a phase of broken symmetry and a region where both phases coexist. The region of coexistence collapses into one of the others via a discontinuous phase transition, whereas the transition between the symmetric phase and the phase of broken symmetry is continuous. The tricritical point where the three regions intersect can be calculated for strong and for weak noise. These limiting values form tight bounds on the tricritical point. In the region of coexistence simulations of finite systems are performed. One finds that the stationary distribution of finite but large systems differs qualitatively from the one of the infinite system. Hence the limits of stationarity and large system size do not commute.

Beta oscillations (13-30 Hz) of the local field potential are a hallmark of Parkinson's disease induced by dopamine deficiency, yet their underlying mechanism remains unclear. This study investigates the complex dynamics underlying beta oscillations of striatal medium spiny neurons in a theoretical model, where the conductance of the M-current (gM) is reduced to reflect dopamine deficiency. For an isolated neuron, codimension-1 and -2 bifurcations underlie excitability transition and the shift of phase response curves (PRCs) from Type 2 to Type 1. In a network with inhibitory coupling, interactions between synchronization regimes and firing frequency of individual neurons generate beta oscillations. Under weak coupling, anti-phase synchronization (AS) changes to global synchronization (GS) via a discontinuous phase transition with a bistable regime. Typical AS and GS are explained by the match between PRC types and inhibitory post-synaptic currents. In the AS, the network frequency is about twice the firing frequency of individual neurons at intermediate gM, resulting in beta network oscillations. As gM decreases further, neurons show higher firing frequencies, and GS replaces AS via a discontinuous transition, yielding beta oscillations with paradoxical frequency reduction. Under strong coupling, partial synchronization emerges instead of AS or GS, accompanied by a continuous transition. The partial synchronization induces beta oscillations across a wide parameter region, where the network frequency exceeds the low firing frequency of individual neurons. These nonlinear dynamics of single neurons and synchronization transitions in the network provide new insights into the generation and modulation of beta oscillations in Parkinson's disease.

The temperature-induced high-spin(HS)-low-spin(LS) transition of the 3d4-3d7 transition-metal ions in crystals (oxides and sulphides) are investigated on a semiquantitative basis by taking into account the elastic interaction between HS (or LS) clusters produced in a LS (or HS) crystal. The critical temperature at which the discontinuous phase transition takes place is evaluated as a function of the ligand-field strength Delta in units of spin-pairing energy Pi . It is shown that the 'crossover range' where this critical temperature exists is restricted to a finite range of Delta / Pi around the crossover point. The crossover ranges are evaluated for the crystals MO, MS2, M2O3, and LaMO3 where M represents transition-metal ions. The temperature variation of the equilibrium concentrations of HS and LS clusters in LS and HS crystals, respectively, is calculated for several values of Delta / Pi in perovskite LaMO3 crystals. It is pointed out that two discontinuous transitions from HS to LS and LS to HS with decreasing temperature may be expected in the d6 and the d7 systems.

We define the characteristic temperature T C of a material by using the behaviour of the specific heat in the vicinity of this temperature: C p,N (T ) = f(1 − (T/T C ),p,N ) for T < T C and C p,N (T ) = g((T/T C ) − 1,p,N ) for T > T C where p is the external pressure and N is the number of particles. The characteristic temperature T C = T C ( p, N ) can e.g. be identified with the transition temperature in the case of continuous phase transitions or with one of the two characteristic temperatures (on heating) and (on cooling) in the case of discontinuous phase transitions. Using the idea of the characteristic temperature, we present a new method on how to calculate the Gibbs potential G = G(T, p, N) in the vicinity of the characteristic temperature, without it being a necessity to know the explicite expressions for f and g. This general and uniform approach allows to derive many thermodynamical relations connecting measurable quantities at the characterisic point as e.g. Clausius–Clapeyron equation and its analogs, valid for discontinuous phase transitions and 15 relations valid for both continuous and discontinuous phase transitions, 6 of them are new. As a special case of these latter relations, we find six relations valid for continuous phase transitions (Ehrenfest equation and its analogs). The derived relations result exclusively from the definition of the characteristic point and are independent from the underlying microscopical mechanism leading to phase transitions (as e.g. solid–liquid, liquid–vapour, magnetic, ferroelectric, structural and superconducting transitions) and therefore, they are generally valid. These relations can be very important for material science. Using some of them it is possible to predict (without performing additional experiments) whether the transition or characteristic temperature T C of a given material sample will increase or decrease under doping (injection) or applied pressure. Similar 24 thermodynamical relations can also be found with ease for systems described by the Helmholtz potential F = F(T,V, N ).

Neutron Diffraction Study of the Low-Temperature Phases of Sodium Oxide Nitrite, Na3ONO2

Based on quasistationary distribution ideas, a general finite size scaling theory is proposed for discontinuous nonequilibrium phase transitions into absorbing states. Analogously to the equilibrium case, we show that quantities such as response functions, cumulants, and equal area probability distributions all scale with the volume, thus allowing proper estimates for the thermodynamic limit. To illustrate these results, five very distinct lattice models displaying nonequilibrium transitions-to single and infinitely many absorbing states-are investigated. The innate difficulties in analyzing absorbing phase transitions are circumvented through quasistationary simulation methods. Our findings (allied to numerical studies in the literature) strongly point to a unifying discontinuous phase transition scaling behavior for equilibrium and this important class of nonequilibrium systems.

Threshold versions of Schloegl's model on a lattice, which involve autocatalytic creation and spontaneous annihilation of particles, can provide a simple prototype for discontinuous non-equilibrium phase transitions. These models are equivalent to so-called threshold contact processes. A discontinuous transition between populated and vacuum states can occur selecting a threshold of N ≥ 2 for the minimum number, N, of neighboring particles enabling autocatalytic creation at an empty site. Fundamental open questions remain given the lack of a thermodynamic framework for analysis. For a square lattice with N = 2, we show that phase coexistence occurs not at a unique value but for a finite range of particle annihilation rate (the natural control parameter). This generic two-phase coexistence also persists when perturbing the model to allow spontaneous particle creation. Such behavior contrasts both the Gibbs phase rule for thermodynamic systems and also previous analysis for this model. We find metastability near the transition corresponding to a non-zero effective line tension, also contrasting previously suggested critical behavior. Mean-field type analysis, extended to treat spatially heterogeneous states, further elucidates model behavior.

In this Chapter, we discuss the effects of higher-order structures on SIS-like processes of social contagion. After a brief motivational introduction where we illustrate the standard SIS process on networks and the difference between simple and complex contagions, we introduce spreading processes on higher-order structures starting from the most general formulation on hypergraphs and then moving to several mean-field and heterogeneous mean-field approaches. The results highlight the rich phenomenology brought by taking into account higher-order contagion effects: both continuous and discontinuous transitions are observed, and critical mass effects emerge. We conclude with a short discussion on the theoretical results regarding the nature of the epidemic transition and the general need for data to validate these models.

There are conflicting reports in the literature regarding the finite-size scaling of the Liouvillian gap and dynamical fluctuations at discontinuous phase transitions, with various studies reporting either exponential or power-law behavior. We clarify this issue by employing large deviation theory. We distinguish two distinct classes of discontinuous phase transitions that have different dynamical properties. The first class is associated with phase coexistence, i.e., the presence of multiple stable attractors of the system dynamics (e.g., local minima of the free-energy functional) in a finite phase diagram region around the phase transition point. In that case, one observes asymptotic exponential scaling related to stochastic switching between attractors (though the onset of exponential scaling may sometimes occur for very large system sizes). In the second class, there is no phase coexistence away from the phase transition point, while at the phase transition point itself there are infinitely many attractors. In that case, one observes power-law scaling related to the diffusive nature of the system relaxation to the stationary state.

A technique based on steady-state fluorescence measurements is introduced to study the discontinuous volume phase transition of hydrogen-bonding gel. The fluorescence light intensity of pyranine (1-Hydroxypyrene-3, 6, 8-trisulfonic acid) bonded to the gel by means of methacryl amino-propyl-trimethyl ammonium chloride was monitored during the swelling of poly(methacrylic acid-co-dimethyl acrylamide) copolymer in water as a function of temperature. A discontinuous volume phase transition around 60 degrees C and an additional phase around 30 degrees C, which may be interpreted as the freely fluctuating phase in a collapsed state, were observed when the fluorescence intensity changes.

In this review, after providing the basic physical concept behind quantum annealing (or adiabatic quantum computation), we present an overview of some recent theoretical as well as experimental developments pointing to the issues which are still debated. With a brief discussion on the fundamental ideas of continuous and discontinuous quantum phase transitions, we discuss the Kibble-Zurek scaling of defect generation following a ramping of a quantum many body system across a quantum critical point. In the process, we discuss associated models, both pure and disordered, and shed light on implementations and some recent applications of the quantum annealing protocols. Furthermore, we discuss the effect of environmental coupling on quantum annealing. Some possible ways to speed up the annealing protocol in closed systems are elaborated upon: we especially focus on the recipes to avoid discontinuous quantum phase transitions occurring in some models where energy gaps vanish exponentially with the system size. This article is part of the theme issue 'Quantum annealing and computation: challenges and perspectives'.