https://arxiv.org/api/qcxMspGo1bXfretYMtt4UCE1oCk 2026-06-10T07:37:21Z 26559 75 15 http://arxiv.org/abs/2606.03429v1 2026-06-02T10:18:36Z

Modeling high-dimensional data is challenging, yet essential to understanding many complex systems. Maximum entropy models such as Ising and Potts models have been used extensively to capture pairwise interactions from correlation patterns in data, allowing to infer graphical representations of complex systems from observations (e.g., from protein sequences or neural population activity). Recently, there has been growing interest in modeling higher-order correlation patterns involving simultaneously three or more variables. While progress has been made in binary data with high-order Ising models, we extend this framework to the more general case of discrete data. We introduce q-state spin models, a complete family of maximum entropy models that generalize the vector Potts model to include long-range and arbitrary high-order interactions. In the pairwise case, our models allow for more diverse interaction types compared to the standard vector Potts model. We discuss their statistical interpretation with examples and relate them to discrete Fourier analysis. Using a loop expansion of the partition function, we show that the statistical properties of spin models are fully captured by the algebraic structure of their interactions. We define gauge transformations under which this structure, and thus the partition function, remains invariant. Models equivalent under gauge transformations can be seen as different representations of the same abstract statistical model, despite generally having interactions of different orders, extending results from the binary case. For practical application to data analysis, we focus on a subset of models known in the binary case as Minimally Complex Models, generalizing them to discrete data. We obtain a closed-form expression for the marginal likelihood of these models, enabling fast model selection. We illustrate their use with simple real-world examples.

2026-06-02T10:18:36Z 89 pages, 16 figures Aaron De Clercq Merijn Moody Clélia de Mulatier http://arxiv.org/abs/2511.02991v3 2026-06-02T09:31:07Z

When liquids are cooled and crystallization is avoided, their dynamics slow dramatically and the material eventually solidifies into an amorphous glass. Experiments show that chemically distinct glass forming liquids share universal features in both the spectral shape and the temperature dependence of the primary structural relaxation. We introduce Randium, a generic, energetically coarse-grained minimal model of viscous liquids. The model, inspired by results from atomistic molecular-dynamics simulations, is implemented on a two-dimensional lattice with Gaussian-distributed nearest-neighbor interactions. Temperature is the only control parameter, and at low temperatures, dynamic facilitation and dynamical heterogeneity emerge from simple nearest-neighbor rearrangements. The relaxation spectra obey time-temperature superposition, and they reproduce shapes observed experimentally for chemically distinct systems. The temperature dependence of the structural-relaxation time follows parabolic scaling, and the relaxation time grows exponentially with the heterogeneity length scale. The absence of elasticity-induced facilitation in Randium shows that this is not required for universal viscous-liquid dynamics. Other explanations for universal relaxation are discussed in light of Randium.

2025-11-04T20:50:28Z 8 pages, 8 figures Ulf R. Pedersen http://arxiv.org/abs/2606.03380v1 2026-06-02T09:24:05Z

Ising Hamiltonians are basic models of disordered magnets and a standard language for quantum and classical optimization. We study an energy-selective quantum search primitive in which the physical evolution \(\exp(-\mathrm{i} T H)\) is used directly as a Hamiltonian phase oracle. Unlike a Boolean oracle, this oracle marks configurations continuously by their phases and selects a finite resonance band rather than a preassigned marked set. We show that alternating it with the Grover diffusion operator nevertheless produces a Grover-type amplification peak. An exact spectral recurrence and a generating-function representation determine the peak position, width, and height. For an annealed Gaussian density of states, target energies in a high-density tail require \(Θ(\sqrt{2^n/M})\) oracle calls when the resonance contains \(M\) configurations. For random Ising spectra, overlap-induced correlations shift and distort the peak; spectral symmetrization and iterative calibration remove this detuning for prescribed-energy targeting.

2026-06-02T09:24:05Z 22 pages, 4 figures A. S. Plyashechnik A. A. Zhukov A. V. Lebedev W. V. Pogosov http://arxiv.org/abs/2606.03300v1 2026-06-02T08:13:40Z

Within the full replica-symmetry-breaking (fullRSB) solution of dense hard spheres in infinite dimension, Charbonneau, Kurchan, Parisi, Urbani, and Zamponi (CKPUZ; J.Stat.Mech.P10009, 2014) introduced three critical exponents $a$, $b$, $c$ governing the matching region of the fullRSB profile near the jamming transition. These exponents satisfy two scaling relations. The first, $b=(1+c)/2$, was established analytically by the diffusion-drift balance in the scaling ansatz. The second, $a+b=1$, was observed numerically to arbitrary precision but could not be proven. The exponents $a,b,c$ of the scaling fullRSB ansatz are related to the physical exponents $α, θ, κ$ that control the gap, force, and overlap distributions by the relations $α=a/b$, $θ=(c-a)/(b-c)$, $κ=c+1$. Crucially, the relation $a+b=1$ yields the scaling relations $α=1/(2+θ)$ and $κ=2-2/(3+θ)$ predicted on independent grounds by the mechanical-marginal-stability arguments of Wyart and collaborators. Here, we give an analytic proof of the identity $a+b=1$ from the scaling fullRSB equations. The proof was obtained through interaction with Claude (Sonnet 4.6 and Opus 4.7) and verified by us.

2026-06-02T08:13:40Z 12 pages, no figures Giorgio Parisi Francesco Zamponi http://arxiv.org/abs/2606.03217v1 2026-06-02T06:23:31Z

Chain-of-thought (CoT) reasoning has become a widely used mechanism for eliciting multi-step reasoning in large language models by generating intermediate reasoning steps at inference time. Yet the scaling behavior of generalization with CoT depth remains poorly understood. To address this question, we study a theoretically solvable model of CoT for in-context weight prediction in linear regression, where test-time reasoning is represented as an iterative refinement of the weight-parameter estimate. Using tools from random matrix theory under high-dimensional asymptotics, we derive an exact formula for the generalization error as a function of reasoning depth, pretraining data amount, and context length. Our analysis reveals a sharp phase transition separating exponential and polynomial improvement, saturation, and overthinking, and characterizes how the optimal reasoning depth scales. We further show that deeper reasoning is most effective with sufficiently rich pretraining and in-context information, whereas limited pretraining or context makes longer reasoning prone to error amplification or saturation. We also validate these predictions through experiments on fully learned linear attention and softmax attention models. Our results provide a unified theoretical account of how test-time CoT depth affects generalization.

2026-06-02T06:23:31Z Kaito Takanami Cengiz Pehlevan http://arxiv.org/abs/2512.06851v3 2026-06-02T03:40:05Z

We investigate reentrant topological transitions in a one-dimensional Su-Schrieffer-Heeger chain with generalized Bernoulli disorder in the intradimer hopping amplitudes. Owing to its independently tunable values and probabilities, the multivalued disorder distribution provides a direct way to control the topological phase diagram. We show that increasing the disorder strength can split the nontrivial regime into multiple disconnected topological windows, whose number, widths, and locations are determined by the distribution parameters. The phase boundaries are derived analytically from the zero-mode inverse localization length and are governed by a weighted geometric mean of the disordered hopping amplitudes, in agreement with numerical results from the reflection-matrix topological quantum number and the real-space winding number. We also show that the mean chiral displacement dynamically identifies these reentrant windows. These results demonstrate how multivalued random disorder can organize and tune reentrant topological behavior in one-dimensional chiral lattices.

2025-12-07T14:05:32Z 15 pages, 12 figures Phys. Rev. A 113, 063502 (2026) Ruijiang Ji Yunbo Zhang Shu Chen Zhihao Xu 10.1103/ynnr-q4h3 http://arxiv.org/abs/2606.02999v1 2026-06-02T01:13:10Z

This study investigates the influence of material density, disorder in onsite energies, and localization length on relaxation dynamics within a two-dimensional random site Coulomb glass model at half-filling. To explore relaxation laws, we calculate the eigenvalue distribution of the linear dynamical matrix using mean-field approximations. Our findings indicate that the system initially undergoes rapid relaxation through energy-lowering transitions. The depletion of the single-particle density of states (DOS) near the Fermi level leads to slow relaxation, with fluctuations diminishing according to a power law. Subsequently, the system adheres to an exponential decay law after a specific period, defined as the relaxation time, which is inversely related to the minimum eigenvalue of the dynamical matrix. As the density of the system decreases, the relaxation rate slows down, resulting in an increase in the relaxation time. For a constant density and localization length, an increase in the disorder of onsite energies results in a longer relaxation time. A significant portion of the eigenvalue spectrum remains unaffected, suggesting that a reduction in localization length concurrent with increased disorder may play an equally vital role in the slow dynamics observed.

2026-06-02T01:13:10Z 8 pages, 11 figures Phys. Rev. E 113, 064107 (2026) Preeti Bhandari Vikas Malik 10.1103/dgph-njm1 http://arxiv.org/abs/2512.20727v2 2026-06-01T21:48:25Z

Organic mixed conductors (OMCs) represent a promising class of materials for applications in bioelectronics, physical computing, and thermoelectrics. Rather unparalleled, OMCs feature dynamics spanning multiple length and time scales, involving an intricate coupling between electronic, ionic, and mass transport. These characteristics set them notably apart from traditional semiconductors and hinder the description by conventional semiconductor theory. In this work, we approach the charge carrier modulation of OMCs using statistical mechanics. We discuss OMCs from a thermodynamic perspective and contrast them with established semiconductor materials, highlighting key differences in their collective charge carrier dynamics. This motivates our description of OMCs as a lattice gas, which we analyze within the grand canonical ensemble. The model exhibits a first-order phase transition analogous to a classical vapor$\unicode{x2013}$liquid transition, governed by temperature and chemical potential. In doing so, it captures the formation of distinct low- and high-density carrier phases, consistent with recently reported experimental observations. It also illustrates how metastability near the phase boundary can give rise to history-dependent characteristics in device operation, a similarly well-reported effect in OMC transistors. This work is intended as a simple motivation for studying OMCs through the lens of statistical mechanics, offering a more natural description than traditional semiconductor models developed for materials of fundamentally distinct character.

2025-12-23T19:36:53Z Lukas M. Bongartz http://arxiv.org/abs/2606.02808v1 2026-06-01T19:27:41Z

We show the emergence of a contribution characteristic of the boson peak in the spectral density of structural glasses. To model the vitreous state, we consider static density-fluctuation fields coupled to a multiplicative quenched disorder. Performing an ensemble average over all disorder realizations, a functional series representation of the average free energy is obtained. In this series representation of the average free energy for the glassy state of matter, we identify in the function space effective actions. These effective actions present a large number of metastable states and ground states. Random first-order transition, widely discussed in the literature as a description of the transition from the supercooled liquid to the glassy state of matter, emerges naturally in our formalism. We establish the connection between the use of hyperbolic differential equations with random coefficients and the presence of many ground states in the average free energy. This connection allows us to study emergent excitations in such amorphous materials.

2026-06-01T19:27:41Z 13 pages, 0 figures M. M. Balbino I. P. de Freitas A. M. S. Macedo G. Krein N. F. Svaiter http://arxiv.org/abs/2606.02794v1 2026-06-01T18:59:01Z

Scaling laws, the power-law relations between loss, architecture size, and compute observed in modern neural networks, offer a quantitative way to characterize the complexity of a learning problem, with the exponent governing the decay of the loss reflecting how rapidly additional resources translate into improved accuracy, and thus how hard the target is to learn. Whether an analogous framework can characterize the complexity of physical problems remains open. We address this question for Neural-Network Quantum States, a leading variational approach for strongly correlated quantum many-body systems. Using transformer wave functions to approximate ground states of the $J_1$-$J_2$ Heisenberg model on triangular and square lattices with up to $20\times 20$ sites, we find that the $V$-score, a measure of accuracy of a variational state, decays as a power law in training compute. Under an appropriate rescaling of compute, results for different system sizes collapse onto a single curve, analogous to scaling collapse in critical phenomena. The resulting power law is, to a good approximation, independent of the number of sites, showing that the transformer Ansatz is size-consistent for the systems considered. The exponent decreases systematically with frustration, identifying it as a quantitative measure of representational difficulty of the ground state and establishing scaling laws as a general framework for benchmarking variational ansätze.

2026-06-01T18:59:01Z 7 pages, 5 figures Riccardo Rende Alessandro Sinibaldi Luciano Loris Viteritti Roeland Wiersema Antoine Georges Giuseppe Carleo http://arxiv.org/abs/2606.02733v1 2026-06-01T18:00:35Z

The interplay between disorder and topology has become a central theme in condensed matter physics. Disorder can not only destroy topological phases but also induce them, as exemplified by the topological Anderson insulator (TAI). Here we show that, in close analogy, disorder can drive the clean-limit, time-reversal-broken(T-broken) quantum spin Hall state of ferromagnetic(FM) monolayer MnBi4Te7 into a quantum anomalous Hall phase, which was called topological Anderson Chern insulator (TACI). Using density functional theory (DFT) and nonequilibrium Green's func tion (NEGF) calculations in the presence of disorder, we identify disorder induced phases-including T-broken TAI, TACI, Normal insulator, etc., then construct a comprehensive phase diagram. To discriminate multiple phases in the strong disorder regime, we further use the density of states computed within the self-consistent Born approximation (SCBA), which in particular distinguishes gapped and ungapped topological phases. We find that the two effective band inversions of Hamiltonian are suppressed at distinct critical disorder strengths; the survival of a single inversion over a finite disorder window stabilizes the TACI. Remarkably, at strong disorder, we further propose a zero Hall plateau insulating state characterized by an insulating bulk and edge channels subject to diffusive scattering that can coexist with the TACI. This behavior is distinct from a conventional band-gap Chern insulator and provides a clear experimental signature.

2026-06-01T18:00:35Z Bo Yin Yan Zhang Anqi Wang Jie Shen Zhijun Wang Quansheng Wu http://arxiv.org/abs/2606.02682v1 2026-06-01T17:51:45Z

In this manuscript, we explore the application of neural networks to predict the natural parameter $κ\geq 0$ of Schramm-Loewner Evolution (SLE$_κ$) theory. SLE$_κ$ is a family of random fractal curves that has significant implications in Statistical Mechanics and Conformal Field Theory. This parameter $κ\geq 0$ plays an important role in the theory as there are models of Planar Statistical Physics that are proven to have SLE as scaling limits as well as others that are conjectured to have this limit for various choices of the parameter $κ\geq 0$. In addition, there are three different statistical behaviors of the SLE curves as the parameter $κ$ changes in $[0, \infty).$ Leveraging the powerful pattern recognition capabilities of neural networks, this study aims to develop a predictive model that can estimate the $κ$ parameter with good accuracy.

2026-06-01T17:51:45Z The version of record which also contains robustness analysis, Deep Learning experiments for the SLE traces, etc. is available at Neural Comput & Applic 38, 318 (2026). https://doi.org/10.1007/s00521-026-12058-3 Neural Comput & Applic 38, 318 (2026) Neilesh Shrotri Vlad Margarint 10.1007/s00521-026-12058-3 http://arxiv.org/abs/2606.02512v1 2026-06-01T17:25:24Z

Geometrical frustration, a central paradigm in condensed matter physics, provides a unifying language for systems in which locally preferred interactions cannot be made globally compatible. Here, we use this language to formulate a minimal theory of frustrated neural timing, mapping repulsively coupled rhythmic units onto antiferromagnetic XY models. Within this framework, the condensed-matter concepts of local constraints, degenerate ground-state manifolds, metastability, and quench dynamics become a concrete diagnostic framework for structured neural phase dynamics. We analyze a hierarchy of geometries: a triangle as the minimal frustrated motif with two chiral 120° timing states, a tetrahedron whose reduced ground-state manifold consists of intersecting continuous branches associated with antipodal pairings, and a kagome lattice on which local constraints define a constrained three-coloring manifold. The kagome lattice reveals the central dynamical result: zero-temperature relaxation suppresses global synchrony but typically selects low-energy metastable torque-balanced states rather than exact ground states. Finally, we show how the phase theory can be carried back towards biophysical neural models by treating it as an effective-interaction target, where geometrical timing frustration is realized through preferred phase lags that become incompatible around closed motifs. This perspective suggests that weak global coherence in neural systems does not necessarily signal disordered activity, but can reflect structured local timing order shaped by a frustrated dynamical landscape.

2026-06-01T17:25:24Z 39 pages, 15 figures Brandon B. Le http://arxiv.org/abs/2606.02476v1 2026-06-01T16:47:48Z

On-site disorder can be leveraged to induce a transition from a trivial to a topological insulator. However it is unclear if structural disorder in the absence of on-site disorder can aid a similar transition and, if so, which kind of structural disorder is more favourable. We numerically show that structural disorder can enhance and sustain a topological phase up to strong disorder in two dimensions provided that one penalises atomic sites from being close to one another. However, we find this effect is absent in three dimensions, where structural disorder appears generically detrimental to the phase. In our calculations we include disorder that can scramble the global spin-reference frame, an overlooked type of disorder expected to exist in strongly disordered solids. This disorder fatally scrambles the information necessary for the spin-Bott and the spin-Chern marker to correctly diagnose a topological phase. By using the spectral localizer, a local marker directly defined using the time-reversal symmetry operator rather than a spin-projection, we show how one can circumvent this limitation, providing a basis-indifferent theory for calculating Z2 invariants. Our work showcases that not all structural disorders are equally beneficial to topology, and highlights guiding principles to enhance and detect topological phases in both solid-state and metamaterial realisations.

2026-06-01T16:47:48Z 19 pages, 13 figures Laura Gomez Paz Peru d'Ornellas Adolfo G Grushin http://arxiv.org/abs/2602.06065v3 2026-06-01T13:46:20Z

Understanding how the structure of language can be learned from sentences alone is a central question in both cognitive science and machine learning. Studies of the internal representations of Large Language Models (LLMs) support their ability to parse text when predicting the next word, while representing semantic notions independently of surface form. Yet, which data statistics make these feats possible, and how much data is required, remain largely unknown. Probabilistic context-free grammars (PCFGs) provide a tractable testbed for studying these questions. However, prior work has focused either on the post-hoc characterization of the parsing-like algorithms used by trained networks; or on the learnability of PCFGs with fixed syntax, where parsing is unnecessary. Here, we (i) introduce a tunable class of PCFGs in which both the degree of ambiguity and the correlation structure across scales can be controlled; (ii) provide a learning mechanism -- an inference algorithm inspired by the structure of deep convolutional networks -- that links learnability and sample complexity to specific language statistics; and (iii) validate our predictions empirically across deep convolutional and transformer-based architectures. Overall, we propose a unifying framework where correlations at different scales lift local ambiguities, enabling the emergence of hierarchical representations of the data.

2026-01-31T17:35:06Z Accepted as regular paper at ICML 2026 Jack T. Parley Francesco Cagnetta Matthieu Wyart