Leïla Moueddene, Nikolaos G Fytas, Bertrand Berche
Since the landmark work of Lee and Yang, locating the zeros of the partition function in the complex magnetic-field plane has become a powerful method for studying phase transitions. Fisher later extended this approach to complex temperatures, and subsequent generalizations introduced other control parameters, such as the crystal field. In previous works [Moueddene et al, J. Stat. Mech. (2024) 023206; Phys. Rev. E 110, 064144 (2024)] we applied this framework to the two- and three-dimensional Blume-Capel model -- a system with a rich phase structure where a second-order critical line meets a first-order line at a tricritical point. We showed that the scaling of Lee-Yang, Fisher, and crystal-field zeros yields accurate critical exponents even for modest lattice sizes. In the present study, we extend this analysis and demonstrate that simulations need not be performed exactly at the nominal transition point to obtain reliable exponent estimates. Strikingly, small system sizes are sufficient, which not only improves methodological efficiency but also advances the broader goal of reducing the carbon footprint of large-scale computational studies.
Abstract This study focuses on the impurities in integrable models from the viewpoint of generalized hydrodynamics (GHD). An impurity can be thought of as a boundary condition for the GHD equation, relating the states on the left and right sides. It was found that, in interacting models, it is not possible to disentangle the incoming and outgoing states, which means that it is not possible to think of scattering as a mapping that maps the incoming state to the outgoing state. A novel class of impurities, dubbed mesoscopic impurities, was then introduced, whose spatial size is mesoscopic (i.e. their size L m i c r o ≪ L i m p ≪ L is much larger than the microscopic length scale L m i c r o but much smaller than the macroscopic scale L). Because of their large size, it is possible to describe mesoscopic impurities via GHD. This simplification allows one to study these impurities both analytically and numerically. These impurities show interesting non-perturbative scattering behavior, such as the nonuniqueness of the solutions and a nonanalytic dependence on the impurity strength. In models with one quasi-particle species and a scattering phase shift that depends only on the difference in momenta, the scattering can be described using an effective Hamiltonian. This Hamiltonian is dressed due to the interaction between the particles and satisfies a self-consistency fixed-point equation. In the example of the hard-rod model, it was demonstrated how this fixed-point equation can be used to find almost explicit solutions to the scattering problem by reducing it to a two-dimensional system of equations that can be solved numerically.
Abstract The well-known Solow growth model is the workhorse model of the theory of economic growth, which studies capital accumulation in a model economy as a function of time with capital stock, labour and technology-based production as the basic ingredients. The capital is assumed to be in the form of manufacturing equipment and materials. Two important parameters of the model are: the saving fraction s of the output of a production function and the technology efficiency parameter A , appearing in the production function. The saved fraction of the output is fully invested in the generation of new capital and the rest is consumed. The capital stock also depreciates as a function of time due to the wearing out of old capital and the increase in the size of the labour population. We propose a stochastic Solow growth model assuming the saving fraction to be a sigmoidal function of the per capita capital k p . We derive analytically the steady state probability distribution P ( k p ) and demonstrate the existence of a poverty trap, of central concern in development economics. In a parameter regime, P ( k p ) is bimodal with the twin peaks corresponding to states of poverty and well-being, respectively. The associated potential landscape has two valleys with fluctuation-driven transitions between them. The mean exit times from the valleys are computed and one finds that the escape from a poverty trap is more favourable at higher values of A . We identify a critical value of A c below (above) which the state of poverty (well-being) dominates and propose two early signatures of the regime shift occurring at A c . The economic model, with conceptual foundations in nonlinear dynamics and statistical mechanics, shares universal features with dynamical models from diverse disciplines like ecology and cell biology.
Abstract We consider the one-dimensional Tonks–Girardeau gas with a space-dependent potential out of equilibrium. We derive the exact dynamics of the system when divided into n boxes and decomposed into energy eigenstates within each box. It is a representation of the wave function that is a mixture between real space and momentum space, with basis elements consisting of plane waves localized in a box, giving rise to the term ‘wavelet’. Using this representation, we derive the emergence of generalized hydrodynamics in appropriate limits without assuming local relaxation. We emphasize that a generalized hydrodynamic behaviour emerges in a high-momentum and short-time limit, in addition to the more common large-space and late-time limit, which is akin to a semi-classical expansion. In this limit, conserved charges do not require numerous particles to be described by generalized hydrodynamics. We also show that this wavelet representation provides an efficient numerical algorithm for a complete description of the out-of-equilibrium dynamics of hardcore bosons.
Shunta Kitahama, Hironobu Yoshida, Ryo Toyota
et al.
We give a rigorous proof of Conjecture 3.1 by Prosen [Prosen T 2010 J. Stat. Mech. $\textbf{2010}$ P07020] on the nilpotent part of the Jordan decomposition of a quadratic fermionic Liouvillian. We also show that the number of the Jordan blocks of each size can be expressed in terms of the coefficients of a polynomial called the $q$-binomial coefficient and describe the procedure to obtain the Jordan canonical form of the nilpotent part.
Abstract We consider the problem of fast time-series data clustering. Building on previous work modeling, the correlation-based Hamiltonian of spin variables we present an updated fast non-expensive agglomerative likelihood clustering algorithm (ALC). The method replaces the optimized genetic algorithm based approach (f-SPC) with an agglomerative recursive merging framework inspired by previous work in econophysics and community detection. The method is tested on noisy synthetic correlated time-series datasets with a built-in cluster structure to demonstrate that the algorithm produces meaningful non-trivial results. We apply it to time-series datasets as large as 20 000 assets and we argue that ALC can reduce computation time costs and resource usage costs for large scale clustering for time-series applications while being serialized, and hence has no obvious parallelization requirement. The algorithm can be an effective choice for state-detection for online learning in a fast non-linear data environment, because the algorithm requires no prior information about the number of clusters.
Abstract We review the state of the art of the problem of heat conduction in one dimensional nonlinear lattices. The peculiar role of finite size and time corrections to the predictions of the hydrodynamic theory is discussed. The emerging scenario indicates that when dealing with systems, whose spatial size is comparable with the mean-free path of peculiar nonlinear excitations, hydrodynamic predictions at leading order are no more predictive. We can conjecture that one should take into account estimates of subleading contributions, which could play a major role in some regions of the parameter space in ‘small’ systems.
Abstract We propose a Langevin equation to describe the quantum Brownian motion of bounded particles based on a distinctive formulation concerning both the fluctuation and dissipation forces. The fluctuation force is similar to that employed in the classical case. It is a white noise with a variance proportional to the temperature. The dissipation force is not restricted to be proportional to the velocity and is determined in a way as to guarantee that the stationary state is given by a density operator of the Gibbs canonical type. To this end we derived an equation that gives the time evolution of the density operator, which turns out to be a quantum Fokker–Planck–Kramers equation. The approach is applied to the harmonic oscillator in which case the dissipation force is found to be non Hermitian and proportional to the velocity and position.
Abstract The random walk with hyperbolic probabilities that we are introducing is an example of stochastic diffusion in a one-dimensional heterogeneous media. Although driven by site-dependent one-step transition probabilities, the process retains some of the features of a simple random walk, shows other traits that one would associate with a biased random walk and, at the same time, presents new properties not related to either of them. In particular, we show how the system is not fully ergodic, as not every statistic can be estimated from a single realization of the process. We further give a geometric interpretation for the origin of these irregular transition probabilities.
Light-cone-like propagation of information is a universal phenomenon of nonequilibrium dynamics of integrable spin systems. In this paper, we investigate propagation of a local impact in the one-dimensional $XY$ model with the anisotropy $γ$ in a magnetic field $h$ by calculating the magnetization profile. Applying a local and instantaneous unitary operation to the ground state, which we refer to as the local-impact protocol, we numerically observe various types of light-cone-like propagation in the parameter region $0\leqγ\leq1$ and $0\leq h \leq2$ of the model. By combining numerical integration with an asymptotic analysis, we find the following: (i) for $|h|\geq|1-γ^{2}|$ except for the case on the line $h=1$ with $0<γ<\sqrt{3}/2$, a wave front propagates with the maximum group velocity of quasiparticles, except for the case $γ=1$ and $0<h<1$, in which there is no clear wave front; (ii) for $|h|<|1-γ^{2}|$ as well as on the line $h=1$ with $0<γ<\sqrt{3}/2$, a second wave front appears owing to multiple local extrema of the group velocity; (iii) for $|h|=|1-γ^{2}|$, edges of the second wave front collapses at the origin, and as a result, the magnetization profile exhibits a ridge at the impacted site. Furthermore, we find by an asymptotic analysis that the height of the wave front decays in a power law in time $t$ with various exponents depending on the model parameters: the wave fronts exhibit a power-law decay $t^{-2/3}$ except for the line $h=1$, on which the decay can be given by either $\sim t^{-3/5}$ or $\sim t^{-1}$; the ridge at the impacted site for $|h|=|1-γ^{2}|$ shows the decay $t^{-1/2}$ as opposed to the decay $t^{-1}$ in other cases.
Abstract An important non-perturbative effect in quantum physics is the energy gap of superconductors, which is exponentially small in the coupling constant. A natural question is whether this effect can be incorporated in the theory of resurgence. In this paper we take some steps in this direction. We conjecture that the perturbative series for the ground state energy of a superconductor is factorially divergent, and that its leading Borel singularity is governed by the superconducting energy gap. We test this conjecture in detail in the attractive Gaudin–Yang model, an exactly solvable model in one dimension with a BCS-like ground state. In order to do this, we develop techniques to calculate the exact perturbative series of its ground state energy up to high order. We also argue that the Borel singularity is of the renormalon type, and we identify a class of diagrams leading to factorial growth. We give additional evidence for the conjecture in other models.
Abstract We define the capacity of a learning machine to be the logarithm of the number (or volume) of the functions it can implement. We review known results, and derive new results, estimating the capacity of several neuronal models: linear and polynomial threshold gates, linear and polynomial threshold gates with constrained weights (binary weights, positive weights), and ReLU neurons. We also derive some capacity estimates and bounds for fully recurrent networks, as well as feedforward networks.
Antoine Maillard, Laura Foini, Alejandro Lage Castellanos
et al.
Improved mean-field technics are a central theme of statistical physics methods applied to inference and learning. We revisit here some of these methods using high-temperature expansions for disordered systems initiated by Plefka, Georges and Yedidia. We derive the Gibbs free entropy and the subsequent self-consistent equations for a generic class of statistical models with correlated matrices and show in particular that many classical approximation schemes, such as adaptive TAP, Expectation-Consistency, or the approximations behind the Vector Approximate Message Passing algorithm all rely on the same assumptions, that are also at the heart of high-temperature expansions. We focus on the case of rotationally invariant random coupling matrices in the `high-dimensional' limit in which the number of samples and the dimension are both large, but with a fixed ratio. This encapsulates many widely studied models, such as Restricted Boltzmann Machines or Generalized Linear Models with correlated data matrices. In this general setting, we show that all the approximation schemes described before are equivalent, and we conjecture that they are exact in the thermodynamic limit in the replica symmetric phases. We achieve this conclusion by resummation of the infinite perturbation series, which generalizes a seminal result of Parisi and Potters. A rigorous derivation of this conjecture is an interesting mathematical challenge. On the way to these conclusions, we uncover several diagrammatical results in connection with free probability and random matrix theory, that are interesting independently of the rest of our work.
Sarah Klein, Cécile Appert-Rolland, Martin R. Evans
We study a two-lane two-species exclusion process inspired by Lin et al. (C. Lin et al. J. Stat. Mech., 2011), that exhibits a non-equilibrium pulsing phase. Particles move on two parallel one-dimensional tracks, with one open and one reflecting boundary. The particle type defines the hopping direction. When only particles hopping towards the open end are allowed to change lane, the system exhibits a phase transition from a low density phase to a pulsing phase depending on the ratio between particle injection and type-changing rate. This phase transition can be observed in the stochastic model as well as in a mean-field description. In the low density phase, the density profile can be predicted analytically. The pulsing phase is characterised by a fast filling of the system and - once filled - by a slowly backwards moving front separating a decreasing dense region and an expanding low density region. The hopping of the front on the discrete lattice is found to create density oscillations, both, in time and space. By means of a stability analysis we can predict the structure of the dense region during the emptying process, characterised by exponentially damped perturbations, both at the open end and near the moving front.