https://arxiv.org/api/XDnojXJy4u2IWue45cQaV2e6Dsg 2026-06-10T16:56:15Z 81779 285 15 http://arxiv.org/abs/2605.29958v1 2026-05-28T14:01:10Z

We extend the ``Brownian bees'' model of Berestycki et al. (2021, 2022) to cooperative reproduction, $kA\to(k{+}1)A$, of a population of $N$ symmetric random walkers with removal, at each birth event, of the particle farthest from the origin. Working in the limit $N\to\infty$, we formulate a hydrodynamic free-boundary problem for this model. Using this formalism, we determine steady state population densities for all~$k$ and prove their linear stability for $k\le 2$ and instability for $k\ge 4$. In the marginal case $k=3$, there is a whole continuous family of steady states at a single, critical ratio of the reproduction and diffusion rates. Above criticality the population undergoes an asymptotically self-similar finite-time collapse to the origin. Below the criticality the population spreads diffusively, but the reproduction remains quantitatively relevant. For $k\ge 4$, the unstable steady state separates regimes of a finite-time collapse and a diffusive spreading. Here the collapse dynamics is asymptotically self-similar, and the population density exhibits a scale separation requiring a matched-asymptotic description. Our analytical predictions are confirmed by numerical solutions of the hydrodynamic free-boundary problem and by Monte Carlo simulations of the original microscopic model.

2026-05-28T14:01:10Z 23 one-column pages, 6 figures Ohad Vilk Baruch Meerson http://arxiv.org/abs/2605.29957v1 2026-05-28T14:00:52Z

Biological, artificial, and physical systems dissipate energy to accurately transmit information. While tools of information theory have been used to characterize information-processing capabilities, how reliably this information is acquired along individual trajectories, and which aspects require a thermodynamic cost, is an open question. In this work, we focus on the stochastic predictability of an arbitrary Langevin dynamics, defined as the pointwise mutual information between the current and future states of a system. We show that the fluctuations of predictability obey a universal thermokinetic decomposition at short times, which reveals that information fluctuations are suppressed by energy dissipation and become stronger with increased dynamical activity. Remarkably, we find that the average predictability, i.e., the short-time mutual information, does not carry any dependence on the underlying thermodynamic and kinetic features. Thus, the role of dissipation at short times is not to enhance information, but to reduce its fluctuations. Such dissipative control is effective only when instantiated by nonlinear operations. Moreover, energy consumption governs short- and long-time precision in stochastic oscillators through structurally different mechanisms that can be independently tuned. Our decomposition offers a fundamental thermodynamic basis for understanding the reliability of information transmission in nonequilibrium systems, the constraints on precision in biological systems, and the design of energy-limited control strategies.

2026-05-28T14:00:52Z Giorgio Nicoletti Daniel M. Busiello http://arxiv.org/abs/2605.29945v1 2026-05-28T13:55:13Z

We compute the low-temperature configurational entropy of a two-dimensional supercooled liquid. Our method, based on a higher-dimensional version of the Grassberger--Procaccia algorithm, can be implemented in a manner that is entirely agnostic with respect to both the dynamics and the theoretical framework, as any genuine notion of order should be. In this construction, entropy is obtained as the decay rate of recurrent structural patterns with increasing patch size, directly linking entropy reduction to the growing persistence of amorphous order. Because the method requires only particle positions, without any knowledge of the interaction potential or even of the particle sizes, it can be applied directly to both equilibrium and nonequilibrium aging configurations. The resulting configurational entropy, together with the higher-order Rényi complexities, agree quantitatively with values obtained from conventional definitions. Remarkably, the entropies measured during aging coincide with their equilibrium counterparts when compared at the same inherent-structure energy.

2026-05-28T13:55:13Z Nina Javerzat Gerhard Jung Jorge Kurchan Misaki Ozawa http://arxiv.org/abs/2605.29884v1 2026-05-28T13:08:14Z

We consider three-dimensional (3D) lattice Abelian Higgs models, with compact U(1) gauge variables coupled to a doubly-charged $N$-component complex scalar field (CLAH). We focus on their phase transitions between the disordered-confined (DC) and ordered-deconfined (OD) phases. When they are continuous they belong to the 3D Abelian Higgs (AH) universality class associated with the stable charged fixed point (CFP) of the renormalization-group flow of the 3D AH field theory, or scalar electrodynamics, describing $N$-component complex scalar fields minimally coupled to a U(1) gauge field. This CFP exists only for a sufficiently large number of components, i.e., $N \ge N_d^*$, where the integer $N_d^*$ depends on the spatial dimension $d$ (for example $N_4^*=183$). To estimate $N_3^*$, we look for the minimum number $N_{\rm cL}$ of scalar components of 3D doubly-charged CLAH models developing continuous transitions along their DC-OD transition line. For this purpose, we present finite-size scaling analyses of Monte Carlo simulations for $N\in[4,10]$, up to lattice sizes $L\approx 100$. The results provide evidence of continuous DC-OD transitions for $N=10$, and weak first-order transitions for $N\le 7$. They are not conclusive for $N=8,\,9$. Therefore, we estimate $N_{\rm cL}=9(1)$.

2026-05-28T13:08:14Z 14 pages Claudio Bonati Filippo Mariani Ettore Vicari http://arxiv.org/abs/2605.29875v1 2026-05-28T12:59:09Z

We perform a large scale simulation of quantum annealing in the Sherrington-Kirkpatrick (SK) spin glass up to a system size $N=40000$ to estimate its ground state energy using the deGennes-Suzuki-Kubo mean-field Ising dynamics, extending the earlier results (reported in Eur. Phys. J. B {\bf 98}, 226 (2025)). Here we numerically solve the deGennes-Suzuki-Kubo annealing dynamics to obtain the spin configurations and subsequently the ground state energy for a given system size at the end of the annealing (to the desired quantum system at the corresponding values of the transverse field), starting from a quantum paramagnetic state. The method shows high efficiency, with an overall algorithmic cost of $O(N^3)$ in estimating the energy of the ground state. We later extend this method to study the ground state energy of the Edwards-Anderson (EA) spin glass on a square lattice.

2026-05-28T12:59:09Z 6 pages, 5 figures Soumyaditya Das Soumyajyoti Biswas Bikas K. Chakrabarti http://arxiv.org/abs/2605.29838v1 2026-05-28T12:20:07Z

We propose gate-parameter Lee-Yang zeros of Loschmidt amplitudes as probes of dynamical phases in finite quantum circuits. We illustrate this approach using a brickwork model, where the time evolution is generated by repeated application of a Floquet operator. The Loschmidt amplitude can be expressed as a rational function of the gate parameters. At fixed system size and large circuit depth, its zeros in one complexified gate parameter, with the other parameter held fixed, condense onto limiting curves. We show that these curves comprise a universal component governed by equimodular Floquet eigenvalues, as described by the Beraha-Kahane-Weiss theorem, together with state-dependent contributions controlled by the overlap of eigenstate of the Floquet operator with the initial state. As one of the parameters is varied, the set of zeros reorganizes abruptly, providing a finite-qubit diagnostic of a dynamical phase transition. This mechanism does not rely on integrability: while integrability enables an exact calculation of the Loschmidt amplitude, the condensation of zeros follows from spectral competition and local unitarity alone.

2026-05-28T12:20:07Z 6 pages Chang Liu Yu Wu Yunfeng Jiang Yang Zhang http://arxiv.org/abs/2602.10637v2 2026-05-28T12:16:20Z

Sampling equilibrium molecular configurations from the Boltzmann distribution is a longstanding challenge. Boltzmann Generators (BGs) address this by combining exact-likelihood generative models with importance sampling, but practical scalability is limited. Meanwhile, coarse-grained surrogates enable the modeling of larger systems by reducing effective dimensionality, yet often lack a reweighting procedure required to ensure asymptotically correct statistics. In this work, we propose Coarse-Grained Boltzmann Generators (CG-BGs), a framework for reduced-order generative modeling with importance sampling in coarse-grained coordinate space. CG-BGs generate samples using a flow-based model and reweight them using a learned potential of mean force (PMF). We show that the PMF can be learned from rapidly converged trajectories via enhanced sampling force matching. Experiments demonstrate that CG-BGs capture solvent-mediated interactions in highly reduced representations while substantially reducing computational cost relative to atomistic BGs, providing a practical route toward equilibrium sampling of larger molecular systems.

2026-02-11T08:37:13Z Accepted at ICML 2026 Weilong Chen Bojun Zhao Jan Eckwert Julija Zavadlav http://arxiv.org/abs/2605.29810v1 2026-05-28T11:55:45Z

The $N$-color Ashkin-Teller model corresponds to $N$ Ising models coupled by four-spin interactions. We consider the two-dimensional case in presence of quenched disorder and use scale invariant scattering theory to determine all the solutions of the exact renormalization group fixed points equations. The weak disorder sector is characterized by a solution that, for any fixed $N$ larger than 1, is a line of fixed points with Ising thermal exponents and continuously varying magnetic exponents. The number of fixed point solutions allowed by the symmetries of the model increases at strong disorder illustrating the growing dependence on the distributions of the two random couplings. The presence of some critical exponents which do not depend on the symmetry parameter $N$ confirms this type of superuniversality as a peculiar feature of random criticality.

2026-05-28T11:55:45Z Youssef Makoudi Gesualdo Delfino http://arxiv.org/abs/2605.29749v1 2026-05-28T10:46:49Z

Identifying subgroups of respondents in psychometric data is traditionally addressed with Latent Class Analysis, which requires the number of classes to be specified a priori and can perform poorly when strong inter-item correlations violate local independence assumptions. We propose a network-theoretic alternative based on community detection in subject-subject similarity networks. To suppress the systematic artifacts induced by the factor structure of the items, the similarity is computed in a low-dimensional factor-score space and the null model for modularity maximisation is obtained by removing the leading (global) mode of the similarity matrix, rather than via the standard Newman--Girvan model. The significance of a detected partition is then assessed against a column-wise resampling null through four complementary observables: the modularity, the differential entropy of the eigenvector point cloud at two neighbourhood scales, and the overlap of the within- and between-community similarity histograms. On a synthetic benchmark with controlled mixture signal, all four metrics correctly identify the homogeneous case as null-compatible -- including the demanding regime of a dataset dominated by a single factor -- and exhibit a graded departure from the null as the cluster separation grows. Applied to 14 widely used psychometric scales, the pipeline isolates a small group of datasets supporting a genuine and directly interpretable modular structure, while the remaining scales fall either in a mixed-signal regime or in one compatible with a single homogeneous community. The significance analysis is independent of the specific community-detection algorithm and provides an operational way to test for modular subject-level structure in questionnaire data.

2026-05-28T10:46:49Z 30 pages, 15 figures Arianna Armanetti NETWORKS research unit, IMT School for Advanced Studies Lucca, Lucca, Italy Luca Cecchetti MOMILAB research group, IMT School for Advanced Studies Lucca, Lucca, Italy Eiko Fried Department of Clinical Psychology, Leiden University, Leiden, The Netherlands Department of Methodology-Statistics Leiden University, Leiden, The Netherlands Diego Garlaschelli NETWORKS research unit, IMT School for Advanced Studies Lucca, Lucca, Italy INdAM-GNAMPA Istituto Nazionale di Alta Matematica `Francesco Severi', Rome Italy Lorentz Institute for Theoretical Physics, University of Leiden, Leiden, The Netherlands Miguel Ibáñez-Berganza NETWORKS research unit, IMT School for Advanced Studies Lucca, Lucca, Italy INdAM-GNAMPA Istituto Nazionale di Alta Matematica `Francesco Severi', Rome Italy http://arxiv.org/abs/2605.29745v1 2026-05-28T10:42:54Z

The Localization Landscape Theory (LLT) offers a classical analogy for understanding Anderson localization through an effective confining potential, whose percolation threshold has been proposed to mark the mobility edge. While this correspondence shows striking numerical agreement in three dimensions, its theoretical foundations remain an open question. In this work, we extend the analysis of the LLT on the Bethe lattice presented in~\cite{Tonetti2026}. In this setting in both the Anderson localization transition and the LLT percolation problem admit exact solutions. Our analysis reveals that the two transitions are distinct, with markedly different critical behaviors. Notably, the LLT percolation transition falls into the standard mean-field universality class, in sharp contrast with the unconventional critical behavior of the Anderson transition on the Bethe lattice. Nonetheless, the LLT framework reproduces several exact results, capturing nontrivial features of the very low-disorder regime: it predicts the position of the isolated eigenvalue, the minimal disorder at which both the LLT percolation curve and the mobility edge first appear, and the Aizenman--Warzel lower bound for localization. We also study the dependence of the LLT percolation threshold on the energy shift, evaluate the LLT prediction for the Density of States, and derive several results on the statistical properties of the variables controlling the problem. Finally, we develop an extreme-value analysis showing that the LLT prediction for the Density of States overestimates the amplitude of the tails close to the boundary of the continuous spectrum. These findings provide an exact analytical benchmark showing that, despite its geometric appeal, the LLT does not generally reproduce the quantum critical properties of Anderson localization, while still offering a powerful tool to understand its very low-disorder regime.

2026-05-28T10:42:54Z 50 pages, 13 figures. arXiv admin note: substantial text overlap with arXiv:2512.04037 Lorenzo Tonetti Leticia F. Cugliandolo Marco Tarzia http://arxiv.org/abs/2605.29739v1 2026-05-28T10:33:45Z

Network psychometrics conceptualises psychological constructs as emergent properties of systems of interacting variables. Energy-based probabilistic models have gained popularity as models of these interactions, but their psychometric application has so far been limited, since most implementations assume binary or ternary responses and rely on limiting inference assumptions. We infer and analyse three generalised spin models of ordinal questionnaire data: the generalised Ising, Blume-Capel (BC), and Blume-Emery-Griffiths (BEG) models. We prove the concavity of the maximum likelihood estimation of the parameters, as well as the gauge invariance of the Ising and BC models. Afterwards, we propose an inference protocol of approximated likelihood maximisation, based on the Monte Carlo estimation of the likelihood gradients. We apply this procedure to eleven psychometric and sociological questionnaires, comparing the inferred spin models against the multivariate Gaussian. We then assess whether the inferred models reproduce the empirical features of the data in terms of principal-component histograms, and histograms of Euclidean and Mahalanobis distances to the mean answer. The multi-modality observed in the histograms of principal components is partially captured by the spin models. This trait of polarisation can be understood, in the light of mean-field theory, as coexistence of stable and metastable phases of the spin models. The BEG model systematically outperforms the other models in capturing the distribution of distances to the mean, while all models underestimate the heavy tails of the Mahalanobis distance. Overall, the analysis witnesses the predictive power of the BEG model, able to account better than others for the abundance of outliers and mean responders, and reveals highly non-linear features of questionnaire data that both Gaussian and spin models fail to account for.

2026-05-28T10:33:45Z 18 pages, 5 figures (plus 28 pages and 7 figures in the appendices) Arianna Armanetti Luca Cecchetti Paolo Sarti Diego Garlaschelli Miguel Ibáñez-Berganza http://arxiv.org/abs/2512.18652v2 2026-05-28T10:30:04Z

Changing environmental conditions can significantly affect the dynamics of disease spread. These changes may arise naturally or result from human interventions; in the latter case, lockdown measures that lead to abrupt but temporary reductions in transmission rates are used to combat disease spread. Yet, the impact of these measures on rare events in heterogeneous populations remains understudied. Here, we analyze the susceptible-infected-susceptible (SIS) model in a stochastic setting where disease extinction -- a sudden clearance of the infection -- occurs via a rare, large fluctuation. We use a semiclassical approximation and numerical simulations on heterogeneous assortative networks, with degree-degree correlations between neighboring nodes, to show how the extinction risk of the disease depends on the lockdown's duration and magnitude, and on the network topology.

2025-12-21T09:00:45Z 10 pages, 7 figures; to appear in Phys. Rev. E (2026) Elad Korngut Michael Assaf http://arxiv.org/abs/2511.07664v2 2026-05-28T09:47:30Z

We investigate the dynamics of the Fermi--Pasta--Ulam--Tsingou chain with long-wavelength random initial data. When the energy per particle is small, thermal equilibrium is not reached on a fast timescale and the system enters prethermalization. The formation of the prethermal state is characterized by the development of a Burgers-type shock and the onset of a turbulent-like spectrum with a time dependent exponent $ζ(t)$ in the inertial range. We perform a significant step forward by demonstrating that these features are robust under generic long-wavelength random initial conditions. By employing advanced probabilistic techniques inspired by the works of Dudley and Talagrand, we derive a sharp asymptotic expression for the average shock time in the thermodynamic limit. For large $p$, this time scales as $(p \sqrt{\log p})^{-1}$, where $p$ is the number of excited modes proving that it is an intensive quantity up to a logarithmic correction in the size of the system.

2025-11-10T22:17:50Z 7 pages + 16 pages of supplemental material Phys. Rev. Lett. 136, 217201 (2026) Matteo Gallone Ricardo Grande Antonio Ponno Stefano Ruffo Erwan Druais 10.1103/qxzs-1t1l http://arxiv.org/abs/2604.21553v2 2026-05-28T09:28:33Z

We study a system of backscattering hard rods in one dimension. Contrary to the usual ballistic hard rods, these hard rods flip the sign of their velocities with a rate $γ$. This leads to the decay of the odd moments of velocity while preserving the even moments: the number of conserved quantities in the system becomes half. The introduction of the flipping rate $γ$ is an integrability-breaking perturbation, and this leads to a change in the transport properties in the system. We show using a Dean-Kawasaki fluctuating hydrodynamic formulation that the unequal space-time correlation of the normal mode phase space densities attains a diffusive form at late times. Also, we show that for $t \gg 1/γ$, the two-time density-density correlation of mass densities spreads in a diffusive manner, and for $t \ll 1/γ$, the correlation spreads ballistically, for a background state given by the Boltzmann distribution. Our results present an elegant framework to study systems where integrability is broken by a stochastic noise.

2026-04-23T11:25:48Z 18 pages, 4 figures Mrinal Jyoti Powdel http://arxiv.org/abs/2602.06791v2 2026-05-28T09:05:59Z

Being probabilistic models, during inference large language models (LLMs) display rare events: behaviour that is far from typical but highly significant. By definition all rare events are hard to see, but the enormous scale of LLM usage means that events completely unobserved during development are likely to become prominent in deployment. Here we present an end-to-end framework for the systematic analysis of rare events in LLMs. We provide a practical implementation spanning theory, efficient generation strategies, probability estimation and error analysis, which we illustrate with concrete examples. We outline extensions and applications to other models and contexts, highlighting the generality of the concepts and techniques presented here.

2026-02-06T15:50:36Z ICML 2026 Oral Spotlight Jake McAllister Dorman Edward Gillman Dominic C. Rose Jamie F. Mair Juan P. Garrahan